{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Converting PyTorch to ONNX" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "True\n" ] } ], "source": [ "import torch\n", "\n", "print(torch.cuda.is_available())" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Defining the model" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "from torch import nn, einsum\n", "import torch.nn.functional as F\n", "\n", "from einops import rearrange\n", "from einops.layers.torch import Rearrange\n", "\n", "# helper methods\n", "\n", "def group_dict_by_key(cond, d):\n", " return_val = [dict(), dict()]\n", " for key in d.keys():\n", " match = bool(cond(key))\n", " ind = int(not match)\n", " return_val[ind][key] = d[key]\n", " return (*return_val,)\n", "\n", "def group_by_key_prefix_and_remove_prefix(prefix, d):\n", " kwargs_with_prefix, kwargs = group_dict_by_key(lambda x: x.startswith(prefix), d)\n", " kwargs_without_prefix = dict(map(lambda x: (x[0][len(prefix):], x[1]), tuple(kwargs_with_prefix.items())))\n", " return kwargs_without_prefix, kwargs\n", "\n", "# classes\n", "\n", "class LayerNorm(nn.Module): # layernorm, but done in the channel dimension #1\n", " def __init__(self, dim, eps = 1e-5):\n", " super().__init__()\n", " self.eps = eps\n", " self.g = nn.Parameter(torch.ones(1, dim, 1, 1))\n", " self.b = nn.Parameter(torch.zeros(1, dim, 1, 1))\n", "\n", " def forward(self, x):\n", " var = torch.var(x, dim = 1, unbiased = False, keepdim = True)\n", " mean = torch.mean(x, dim = 1, keepdim = True)\n", " return (x - mean) / (var + self.eps).sqrt() * self.g + self.b\n", "\n", "class PreNorm(nn.Module):\n", " def __init__(self, dim, fn):\n", " super().__init__()\n", " self.norm = LayerNorm(dim)\n", " self.fn = fn\n", " def forward(self, x, **kwargs):\n", " x = self.norm(x)\n", " return self.fn(x, **kwargs)\n", "\n", "class FeedForward(nn.Module):\n", " def __init__(self, dim, mult = 4, dropout = 0.):\n", " super().__init__()\n", " self.net = nn.Sequential(\n", " nn.Conv2d(dim, dim * mult, 1),\n", " nn.GELU(),\n", " nn.Dropout(dropout),\n", " nn.Conv2d(dim * mult, dim, 1),\n", " nn.Dropout(dropout)\n", " )\n", " def forward(self, x):\n", " return self.net(x)\n", "\n", "class DepthWiseConv2d(nn.Module):\n", " def __init__(self, dim_in, dim_out, kernel_size, padding, stride, bias = True):\n", " super().__init__()\n", " self.net = nn.Sequential(\n", " nn.Conv2d(dim_in, dim_in, kernel_size = kernel_size, padding = padding, groups = dim_in, stride = stride, bias = bias),\n", " nn.BatchNorm2d(dim_in),\n", " nn.Conv2d(dim_in, dim_out, kernel_size = 1, bias = bias)\n", " )\n", " def forward(self, x):\n", " return self.net(x)\n", "\n", "class Attention(nn.Module):\n", " def __init__(self, dim, proj_kernel, kv_proj_stride, heads = 8, dim_head = 64, dropout = 0.):\n", " super().__init__()\n", " inner_dim = dim_head * heads\n", " padding = proj_kernel // 2\n", " self.heads = heads\n", " self.scale = dim_head ** -0.5\n", "\n", " self.attend = nn.Softmax(dim = -1)\n", " self.dropout = nn.Dropout(dropout)\n", "\n", " self.to_q = DepthWiseConv2d(dim, inner_dim, proj_kernel, padding = padding, stride = 1, bias = False)\n", " self.to_kv = DepthWiseConv2d(dim, inner_dim * 2, proj_kernel, padding = padding, stride = kv_proj_stride, bias = False)\n", "\n", " self.to_out = nn.Sequential(\n", " nn.Conv2d(inner_dim, dim, 1),\n", " nn.Dropout(dropout)\n", " )\n", "\n", " def forward(self, x):\n", " shape = x.shape\n", " b, n, _, y, h = *shape, self.heads\n", " q, k, v = (self.to_q(x), *self.to_kv(x).chunk(2, dim = 1))\n", " q, k, v = map(lambda t: rearrange(t, 'b (h d) x y -> (b h) (x y) d', h = h), (q, k, v))\n", "\n", " dots = einsum('b i d, b j d -> b i j', q, k) * self.scale\n", "\n", " attn = self.attend(dots)\n", " attn = self.dropout(attn)\n", "\n", " out = einsum('b i j, b j d -> b i d', attn, v)\n", " out = rearrange(out, '(b h) (x y) d -> b (h d) x y', h = h, y = y)\n", " return self.to_out(out)\n", "\n", "class Transformer(nn.Module):\n", " def __init__(self, dim, proj_kernel, kv_proj_stride, depth, heads, dim_head = 64, mlp_mult = 4, dropout = 0.):\n", " super().__init__()\n", " self.layers = nn.ModuleList([])\n", " for _ in range(depth):\n", " self.layers.append(nn.ModuleList([\n", " PreNorm(dim, Attention(dim, proj_kernel = proj_kernel, kv_proj_stride = kv_proj_stride, heads = heads, dim_head = dim_head, dropout = dropout)),\n", " PreNorm(dim, FeedForward(dim, mlp_mult, dropout = dropout))\n", " ]))\n", " def forward(self, x):\n", " for attn, ff in self.layers:\n", " x = attn(x) + x\n", " x = ff(x) + x\n", " return x\n", "\n", "class CvT(nn.Module):\n", " def __init__(\n", " self,\n", " *,\n", " num_classes,\n", " s1_emb_dim = 64,\n", " s1_emb_kernel = 7,\n", " s1_emb_stride = 4,\n", " s1_proj_kernel = 3,\n", " s1_kv_proj_stride = 2,\n", " s1_heads = 1,\n", " s1_depth = 1,\n", " s1_mlp_mult = 4,\n", " s2_emb_dim = 192,\n", " s2_emb_kernel = 3,\n", " s2_emb_stride = 2,\n", " s2_proj_kernel = 3,\n", " s2_kv_proj_stride = 2,\n", " s2_heads = 3,\n", " s2_depth = 2,\n", " s2_mlp_mult = 4,\n", " s3_emb_dim = 384,\n", " s3_emb_kernel = 3,\n", " s3_emb_stride = 2,\n", " s3_proj_kernel = 3,\n", " s3_kv_proj_stride = 2,\n", " s3_heads = 6,\n", " s3_depth = 10,\n", " s3_mlp_mult = 4,\n", " dropout = 0.\n", " ):\n", " super().__init__()\n", " kwargs = dict(locals())\n", "\n", " dim = 1\n", " layers = []\n", "\n", " for prefix in ('s1', 's2', 's3'):\n", " config, kwargs = group_by_key_prefix_and_remove_prefix(f'{prefix}_', kwargs)\n", "\n", " layers.append(nn.Sequential(\n", " nn.Conv2d(dim, config['emb_dim'], kernel_size = config['emb_kernel'], padding = (config['emb_kernel'] // 2), stride = config['emb_stride']),\n", " LayerNorm(config['emb_dim']),\n", " Transformer(dim = config['emb_dim'], proj_kernel = config['proj_kernel'], kv_proj_stride = config['kv_proj_stride'], depth = config['depth'], heads = config['heads'], mlp_mult = config['mlp_mult'], dropout = dropout)\n", " ))\n", "\n", " dim = config['emb_dim']\n", "\n", " self.layers = nn.Sequential(*layers)\n", "\n", " self.to_logits = nn.Sequential(\n", " nn.AdaptiveAvgPool2d(1),\n", " Rearrange('... () () -> ...'),\n", " nn.Linear(dim, num_classes)\n", " )\n", "\n", " def forward(self, x):\n", " latents = self.layers(x)\n", " return self.to_logits(latents)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Loading the model" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": [ "model = CvT(num_classes = 2)\n", "model.load_state_dict(torch.load('../model/test-5-cvt-model.pth'))\n", "\n", "dummy_input = torch.randn(1, 1, 256, 256)\n", "\n", "model.eval()\n", "torch_out = model(dummy_input)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Converting to ONNX" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:204: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor').\n", " inferred_length: int = length // known_product\n", "/home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:316: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!\n", " known: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] != _unknown_axis_length}\n", "/home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:317: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!\n", " unknown: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] == _unknown_axis_length}\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "graph(%input : Float(*, 1, 256, 256, strides=[65536, 65536, 256, 1], requires_grad=0, device=cpu),\n", " %layers.0.0.weight : Float(64, 1, 7, 7, strides=[49, 49, 7, 1], requires_grad=1, device=cpu),\n", " %layers.0.0.bias : Float(64, strides=[1], requires_grad=1, device=cpu),\n", " %layers.0.1.g : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.1.b : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.norm.g : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.norm.b : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.fn.to_q.net.2.weight : Float(64, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.fn.to_kv.net.2.weight : Float(128, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.fn.to_out.0.weight : Float(64, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.0.fn.to_out.0.bias : Float(64, strides=[1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.norm.g : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.norm.b : Float(1, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.fn.net.0.weight : Float(256, 64, 1, 1, strides=[64, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.fn.net.0.bias : Float(256, strides=[1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.fn.net.3.weight : Float(64, 256, 1, 1, strides=[256, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.0.2.layers.0.1.fn.net.3.bias : Float(64, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.0.weight : Float(192, 64, 3, 3, strides=[576, 9, 3, 1], requires_grad=1, device=cpu),\n", " %layers.1.0.bias : Float(192, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.1.g : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.1.b : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.norm.g : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.norm.b : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.fn.to_q.net.2.weight : Float(192, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.fn.to_kv.net.2.weight : Float(384, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.fn.to_out.0.weight : Float(192, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.0.fn.to_out.0.bias : Float(192, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.norm.g : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.norm.b : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.fn.net.0.weight : Float(768, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.fn.net.0.bias : Float(768, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.fn.net.3.weight : Float(192, 768, 1, 1, strides=[768, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.0.1.fn.net.3.bias : Float(192, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.norm.g : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.norm.b : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.fn.to_q.net.2.weight : Float(192, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.fn.to_kv.net.2.weight : Float(384, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.fn.to_out.0.weight : Float(192, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.0.fn.to_out.0.bias : Float(192, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.norm.g : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.norm.b : Float(1, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.fn.net.0.weight : Float(768, 192, 1, 1, strides=[192, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.fn.net.0.bias : Float(768, strides=[1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.fn.net.3.weight : Float(192, 768, 1, 1, strides=[768, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.1.2.layers.1.1.fn.net.3.bias : Float(192, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.0.weight : Float(384, 192, 3, 3, strides=[1728, 9, 3, 1], requires_grad=1, device=cpu),\n", " %layers.2.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.1.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.1.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.0.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.1.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.2.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.3.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.4.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.5.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.6.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.7.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.8.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.fn.to_q.net.2.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.fn.to_kv.net.2.weight : Float(768, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.fn.to_out.0.weight : Float(384, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.0.fn.to_out.0.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.norm.g : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.norm.b : Float(1, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.fn.net.0.weight : Float(1536, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.fn.net.0.bias : Float(1536, strides=[1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.fn.net.3.weight : Float(384, 1536, 1, 1, strides=[1536, 1, 1, 1], requires_grad=1, device=cpu),\n", " %layers.2.2.layers.9.1.fn.net.3.bias : Float(384, strides=[1], requires_grad=1, device=cpu),\n", " %to_logits.2.weight : Float(2, 384, strides=[384, 1], requires_grad=1, device=cpu),\n", " %to_logits.2.bias : Float(2, strides=[1], requires_grad=1, device=cpu),\n", " %4052 : Float(64, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4053 : Float(64, strides=[1], requires_grad=0, device=cpu),\n", " %4055 : Float(64, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4056 : Float(64, strides=[1], requires_grad=0, device=cpu),\n", " %4058 : Float(192, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4059 : Float(192, strides=[1], requires_grad=0, device=cpu),\n", " %4061 : Float(192, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4062 : Float(192, strides=[1], requires_grad=0, device=cpu),\n", " %4064 : Float(192, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4065 : Float(192, strides=[1], requires_grad=0, device=cpu),\n", " %4067 : Float(192, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4068 : Float(192, strides=[1], requires_grad=0, device=cpu),\n", " %4070 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4071 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4073 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4074 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4076 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4077 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4079 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4080 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4082 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4083 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4085 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4086 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4088 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4089 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4091 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4092 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4094 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4095 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4097 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4098 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4100 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4101 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4103 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4104 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4106 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4107 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4109 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4110 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4112 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4113 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4115 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4116 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4118 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4119 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4121 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4122 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4124 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4125 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4127 : Float(384, 1, 3, 3, strides=[9, 9, 3, 1], requires_grad=0, device=cpu),\n", " %4128 : Float(384, strides=[1], requires_grad=0, device=cpu),\n", " %4129 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4130 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4131 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4132 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4133 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4134 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4135 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4136 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4137 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4138 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4139 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4140 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4141 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4142 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4143 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4144 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4145 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4146 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4147 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4148 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4149 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4150 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4151 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4152 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4153 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4154 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4155 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4156 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4157 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4158 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4159 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4160 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4161 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4162 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4163 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4164 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4165 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4166 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4167 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4168 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4169 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4170 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4171 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4172 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4173 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4174 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4175 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4176 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4177 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4178 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4179 : Long(1, strides=[1], requires_grad=0, device=cpu),\n", " %4180 : Long(1, strides=[1], requires_grad=0, device=cpu)):\n", " %327 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[7, 7], pads=[3, 3, 3, 3], strides=[4, 4]](%input, %layers.0.0.weight, %layers.0.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %328 : Float(*, 1, 64, 64, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%327) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %329 : Float(*, 64, 64, 64, device=cpu) = onnx::Sub(%327, %328) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %330 : Float(*, 64, 64, 64, device=cpu) = onnx::Mul(%329, %329) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %331 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%330) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %332 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%327) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %333 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sub(%327, %332) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %334 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %335 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%331, %334) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %336 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%335) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %337 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Div(%333, %336) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %338 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%337, %layers.0.1.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %339 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%338, %layers.0.1.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %340 : Float(*, 1, 64, 64, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%339) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %341 : Float(*, 64, 64, 64, device=cpu) = onnx::Sub(%339, %340) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %342 : Float(*, 64, 64, 64, device=cpu) = onnx::Mul(%341, %341) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %343 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%342) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %344 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%339) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %345 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sub(%339, %344) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %346 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %347 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%343, %346) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %348 : Float(*, 1, 64, 64, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%347) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %349 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Div(%345, %348) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %350 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%349, %layers.0.2.layers.0.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.1 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%350, %layers.0.2.layers.0.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %352 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.1) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %353 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %354 : Long(device=cpu) = onnx::Gather[axis=0](%352, %353) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.8 : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=64, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.1, %4052, %4053) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor : Float(*, 64, 64, 64, strides=[262144, 4096, 64, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.8, %layers.0.2.layers.0.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.16 : Float(*, 64, 32, 32, strides=[65536, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=64, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.1, %4055, %4056) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %360 : Float(*, 128, 32, 32, strides=[131072, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.16, %layers.0.2.layers.0.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %361 : Long(4, strides=[1], device=cpu) = onnx::Shape(%360) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %362 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %363 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%361, %362) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %364 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %365 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %366 : Long(1, strides=[1], device=cpu) = onnx::Add(%363, %365) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %367 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %368 : Long(1, strides=[1], device=cpu) = onnx::Div(%366, %367) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %369 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %370 : Long(1, strides=[1], device=cpu) = onnx::Mul(%368, %369) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %371 : Float(*, *, *, *, strides=[131072, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Slice(%360, %364, %370, %362) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %372 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %373 : Long(1, strides=[1], device=cpu) = onnx::Mul(%368, %372) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %374 : Float(*, *, *, *, strides=[131072, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Slice(%360, %370, %373, %362) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %375 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %376 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %377 : Long(device=cpu) = onnx::Gather[axis=0](%375, %376) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %378 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %379 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %380 : Long(device=cpu) = onnx::Gather[axis=0](%378, %379) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %381 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %382 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %383 : Long(device=cpu) = onnx::Gather[axis=0](%381, %382) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %384 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %385 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %386 : Long(device=cpu) = onnx::Gather[axis=0](%384, %385) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %387 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %388 : Long(device=cpu) = onnx::Div(%377, %387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %389 : Long(device=cpu) = onnx::Cast[to=7](%388) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %390 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%389) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %391 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %392 : Long(device=cpu) = onnx::Div(%380, %391) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %393 : Long(device=cpu) = onnx::Cast[to=7](%392) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %394 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%393) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %395 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %396 : Long(device=cpu) = onnx::Div(%383, %395) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %397 : Long(device=cpu) = onnx::Cast[to=7](%396) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %398 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%397) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %399 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %400 : Long(device=cpu) = onnx::Div(%386, %399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %401 : Long(device=cpu) = onnx::Cast[to=7](%400) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%401) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %403 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %404 : Long(requires_grad=0, device=cpu) = onnx::Mul(%390, %403) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %405 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %406 : Long(requires_grad=0, device=cpu) = onnx::Mul(%404, %405) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %407 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %408 : Long(requires_grad=0, device=cpu) = onnx::Mul(%398, %407) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %409 : Long(requires_grad=0, device=cpu) = onnx::Mul(%408, %inferred_length) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %411 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %412 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%390, %411)\n", " %415 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %416 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%394, %415)\n", " %417 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %418 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%398, %417)\n", " %419 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %420 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length, %419)\n", " %421 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%412, %4129, %416, %418, %420) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.3 : Float(*, *, *, *, *, strides=[262144, 262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor, %421) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.7 : Float(*, *, *, *, *, strides=[262144, 262144, 64, 1, 4096], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.3) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %424 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %425 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%406, %424)\n", " %426 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %427 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%409, %426)\n", " %428 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %429 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%394, %428)\n", " %430 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%425, %427, %429) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q : Float(*, *, *, strides=[4096, 1, 4096], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.7, %430) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %432 : Long(4, strides=[1], device=cpu) = onnx::Shape(%371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %433 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %434 : Long(device=cpu) = onnx::Gather[axis=0](%432, %433) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %435 : Long(4, strides=[1], device=cpu) = onnx::Shape(%371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %436 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %437 : Long(device=cpu) = onnx::Gather[axis=0](%435, %436) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %438 : Long(4, strides=[1], device=cpu) = onnx::Shape(%371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %439 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %440 : Long(device=cpu) = onnx::Gather[axis=0](%438, %439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %441 : Long(4, strides=[1], device=cpu) = onnx::Shape(%371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %442 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %443 : Long(device=cpu) = onnx::Gather[axis=0](%441, %442) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %444 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %445 : Long(device=cpu) = onnx::Div(%434, %444) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %446 : Long(device=cpu) = onnx::Cast[to=7](%445) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %447 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%446) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %448 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %449 : Long(device=cpu) = onnx::Div(%437, %448) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %450 : Long(device=cpu) = onnx::Cast[to=7](%449) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %451 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%450) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %452 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %453 : Long(device=cpu) = onnx::Div(%440, %452) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %454 : Long(device=cpu) = onnx::Cast[to=7](%453) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %455 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%454) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %456 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %457 : Long(device=cpu) = onnx::Div(%443, %456) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %458 : Long(device=cpu) = onnx::Cast[to=7](%457) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.3 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%458) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %460 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %461 : Long(requires_grad=0, device=cpu) = onnx::Mul(%447, %460) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %462 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %463 : Long(requires_grad=0, device=cpu) = onnx::Mul(%461, %462) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %464 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %465 : Long(requires_grad=0, device=cpu) = onnx::Mul(%455, %464) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %466 : Long(requires_grad=0, device=cpu) = onnx::Mul(%465, %inferred_length.3) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %468 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %469 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%447, %468)\n", " %472 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %473 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%451, %472)\n", " %474 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %475 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%455, %474)\n", " %476 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %477 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.3, %476)\n", " %478 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%469, %4130, %473, %475, %477) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.11 : Float(*, *, *, *, *, strides=[65536, 65536, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%371, %478) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.15 : Float(*, *, *, *, *, strides=[65536, 65536, 32, 1, 1024], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.11) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %481 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %482 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%463, %481)\n", " %483 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %484 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%466, %483)\n", " %485 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %486 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%451, %485)\n", " %487 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%482, %484, %486) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k : Float(*, *, *, strides=[1024, 1, 1024], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.15, %487) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %489 : Long(4, strides=[1], device=cpu) = onnx::Shape(%374) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %490 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %491 : Long(device=cpu) = onnx::Gather[axis=0](%489, %490) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %492 : Long(4, strides=[1], device=cpu) = onnx::Shape(%374) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %493 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %494 : Long(device=cpu) = onnx::Gather[axis=0](%492, %493) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %495 : Long(4, strides=[1], device=cpu) = onnx::Shape(%374) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %496 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %497 : Long(device=cpu) = onnx::Gather[axis=0](%495, %496) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %498 : Long(4, strides=[1], device=cpu) = onnx::Shape(%374) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %499 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %500 : Long(device=cpu) = onnx::Gather[axis=0](%498, %499) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %501 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %502 : Long(device=cpu) = onnx::Div(%491, %501) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %503 : Long(device=cpu) = onnx::Cast[to=7](%502) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %504 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%503) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %505 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %506 : Long(device=cpu) = onnx::Div(%494, %505) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %507 : Long(device=cpu) = onnx::Cast[to=7](%506) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %508 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%507) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %509 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %510 : Long(device=cpu) = onnx::Div(%497, %509) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %511 : Long(device=cpu) = onnx::Cast[to=7](%510) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %512 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%511) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %513 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %514 : Long(device=cpu) = onnx::Div(%500, %513) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %515 : Long(device=cpu) = onnx::Cast[to=7](%514) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.7 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%515) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %517 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %518 : Long(requires_grad=0, device=cpu) = onnx::Mul(%504, %517) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %519 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %520 : Long(requires_grad=0, device=cpu) = onnx::Mul(%518, %519) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %521 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %522 : Long(requires_grad=0, device=cpu) = onnx::Mul(%512, %521) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %523 : Long(requires_grad=0, device=cpu) = onnx::Mul(%522, %inferred_length.7) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %525 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %526 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%504, %525)\n", " %529 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %530 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%508, %529)\n", " %531 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %532 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%512, %531)\n", " %533 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %534 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.7, %533)\n", " %535 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%526, %4131, %530, %532, %534) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.19 : Float(*, *, *, *, *, strides=[65536, 65536, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%374, %535) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.23 : Float(*, *, *, *, *, strides=[65536, 65536, 32, 1, 1024], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.19) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %538 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %539 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%520, %538)\n", " %540 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %541 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%523, %540)\n", " %542 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %543 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%508, %542)\n", " %544 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%539, %541, %543) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v : Float(*, *, *, strides=[1024, 1, 1024], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.23, %544) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %546 : Float(*, *, *, strides=[4194304, 1024, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q, %k) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %547 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.20 : Float(*, *, *, strides=[4194304, 1024, 1], requires_grad=1, device=cpu) = onnx::Mul(%546, %547) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.24 : Float(*, *, *, strides=[4194304, 1024, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.20) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.27 : Float(*, *, *, strides=[262144, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.24, %v) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %551 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.27) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %552 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %553 : Long(device=cpu) = onnx::Gather[axis=0](%551, %552) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %554 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.27) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %555 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %556 : Long(device=cpu) = onnx::Gather[axis=0](%554, %555) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %557 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.27) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %558 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %559 : Long(device=cpu) = onnx::Gather[axis=0](%557, %558) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %560 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %561 : Long(device=cpu) = onnx::Div(%553, %560) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %562 : Long(device=cpu) = onnx::Cast[to=7](%561) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %563 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%562) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %564 : Long(device=cpu) = onnx::Div(%556, %354) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %565 : Long(device=cpu) = onnx::Cast[to=7](%564) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %566 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%565) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %567 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %568 : Long(device=cpu) = onnx::Div(%559, %567) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %569 : Long(device=cpu) = onnx::Cast[to=7](%568) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.11 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%569) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %572 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %573 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%563, %572)\n", " %576 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %577 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%566, %576)\n", " %578 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %579 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%354, %578)\n", " %580 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %581 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.11, %580)\n", " %582 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%573, %4132, %577, %579, %581) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.31 : Float(*, *, *, *, *, strides=[262144, 262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.27, %582) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.35 : Float(*, *, *, *, *, strides=[262144, 262144, 1, 4096, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.31) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %585 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %586 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%563, %585)\n", " %587 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %588 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.11, %587)\n", " %589 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %590 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%566, %589)\n", " %591 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %592 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%354, %591)\n", " %593 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%586, %588, %590, %592) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.28 : Float(*, *, *, *, strides=[64, 1, 4096, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.35, %593) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.32 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.28, %layers.0.2.layers.0.0.fn.to_out.0.weight, %layers.0.2.layers.0.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %596 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%input.32, %339) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %597 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%596) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %598 : Float(*, 64, *, *, device=cpu) = onnx::Sub(%596, %597) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %599 : Float(*, 64, *, *, device=cpu) = onnx::Mul(%598, %598) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %600 : Float(*, 1, *, *, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%599) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %601 : Float(*, 1, *, *, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%596) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %602 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sub(%596, %601) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %603 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %604 : Float(*, 1, *, *, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%600, %603) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %605 : Float(*, 1, *, *, strides=[4096, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%604) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %606 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Div(%602, %605) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %607 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%606, %layers.0.2.layers.0.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.36 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%607, %layers.0.2.layers.0.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %609 : Float(*, 256, *, *, strides=[1048576, 4096, 64, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.36, %layers.0.2.layers.0.1.fn.net.0.weight, %layers.0.2.layers.0.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %610 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %611 : Float(*, 256, *, *, device=cpu) = onnx::Div(%609, %610) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %612 : Float(*, 256, *, *, device=cpu) = onnx::Erf(%611) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %613 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %614 : Float(*, 256, *, *, device=cpu) = onnx::Add(%612, %613) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %615 : Float(*, 256, *, *, device=cpu) = onnx::Mul(%609, %614) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %616 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.40 : Float(*, 256, *, *, strides=[1048576, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%615, %616) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.44 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.40, %layers.0.2.layers.0.1.fn.net.3.weight, %layers.0.2.layers.0.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.48 : Float(*, 64, *, *, strides=[262144, 4096, 64, 1], requires_grad=1, device=cpu) = onnx::Add(%input.44, %596) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %620 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.48, %layers.1.0.weight, %layers.1.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %621 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%620) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %622 : Float(*, 192, *, *, device=cpu) = onnx::Sub(%620, %621) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %623 : Float(*, 192, *, *, device=cpu) = onnx::Mul(%622, %622) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %624 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%623) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %625 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%620) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %626 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sub(%620, %625) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %627 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %628 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%624, %627) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %629 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%628) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %630 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Div(%626, %629) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %631 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%630, %layers.1.1.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %632 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%631, %layers.1.1.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %633 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%632) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %634 : Float(*, 192, *, *, device=cpu) = onnx::Sub(%632, %633) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %635 : Float(*, 192, *, *, device=cpu) = onnx::Mul(%634, %634) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %636 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%635) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %637 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%632) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %638 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sub(%632, %637) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %639 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %640 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%636, %639) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %641 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%640) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %642 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Div(%638, %641) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %643 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%642, %layers.1.2.layers.0.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.52 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%643, %layers.1.2.layers.0.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %645 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.52) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %646 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %647 : Long(device=cpu) = onnx::Gather[axis=0](%645, %646) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.60 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=192, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.52, %4058, %4059) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.39 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.60, %layers.1.2.layers.0.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.68 : Float(*, 192, *, *, strides=[49152, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=192, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.52, %4061, %4062) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %653 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.68, %layers.1.2.layers.0.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %654 : Long(4, strides=[1], device=cpu) = onnx::Shape(%653) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %655 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %656 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%654, %655) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %657 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %658 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %659 : Long(1, strides=[1], device=cpu) = onnx::Add(%656, %658) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %660 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %661 : Long(1, strides=[1], device=cpu) = onnx::Div(%659, %660) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %662 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %663 : Long(1, strides=[1], device=cpu) = onnx::Mul(%661, %662) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %664 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Slice(%653, %657, %663, %655) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %665 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %666 : Long(1, strides=[1], device=cpu) = onnx::Mul(%661, %665) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %667 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Slice(%653, %663, %666, %655) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %668 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %669 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %670 : Long(device=cpu) = onnx::Gather[axis=0](%668, %669) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %671 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %672 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %673 : Long(device=cpu) = onnx::Gather[axis=0](%671, %672) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %674 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %675 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %676 : Long(device=cpu) = onnx::Gather[axis=0](%674, %675) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %677 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %678 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %679 : Long(device=cpu) = onnx::Gather[axis=0](%677, %678) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %680 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %681 : Long(device=cpu) = onnx::Div(%670, %680) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %682 : Long(device=cpu) = onnx::Cast[to=7](%681) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %683 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%682) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %684 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %685 : Long(device=cpu) = onnx::Div(%673, %684) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %686 : Long(device=cpu) = onnx::Cast[to=7](%685) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %687 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%686) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %688 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %689 : Long(device=cpu) = onnx::Div(%676, %688) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %690 : Long(device=cpu) = onnx::Cast[to=7](%689) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %691 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%690) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %692 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %693 : Long(device=cpu) = onnx::Div(%679, %692) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %694 : Long(device=cpu) = onnx::Cast[to=7](%693) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.15 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%694) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %696 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %697 : Long(requires_grad=0, device=cpu) = onnx::Mul(%683, %696) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %698 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %699 : Long(requires_grad=0, device=cpu) = onnx::Mul(%697, %698) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %700 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %701 : Long(requires_grad=0, device=cpu) = onnx::Mul(%691, %700) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %702 : Long(requires_grad=0, device=cpu) = onnx::Mul(%701, %inferred_length.15) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %704 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %705 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%683, %704)\n", " %708 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %709 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%687, %708)\n", " %710 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %711 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%691, %710)\n", " %712 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %713 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.15, %712)\n", " %714 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%705, %4133, %709, %711, %713) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.43 : Float(*, *, *, *, *, strides=[196608, 65536, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.39, %714) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.47 : Float(*, *, *, *, *, strides=[196608, 65536, 32, 1, 1024], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.43) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %717 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %718 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%699, %717)\n", " %719 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %720 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%702, %719)\n", " %721 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %722 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%687, %721)\n", " %723 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%718, %720, %722) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.3 : Float(*, *, *, strides=[65536, 1, 1024], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.47, %723) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %725 : Long(4, strides=[1], device=cpu) = onnx::Shape(%664) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %726 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %727 : Long(device=cpu) = onnx::Gather[axis=0](%725, %726) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %728 : Long(4, strides=[1], device=cpu) = onnx::Shape(%664) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %729 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %730 : Long(device=cpu) = onnx::Gather[axis=0](%728, %729) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %731 : Long(4, strides=[1], device=cpu) = onnx::Shape(%664) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %732 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %733 : Long(device=cpu) = onnx::Gather[axis=0](%731, %732) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %734 : Long(4, strides=[1], device=cpu) = onnx::Shape(%664) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %735 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %736 : Long(device=cpu) = onnx::Gather[axis=0](%734, %735) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %737 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %738 : Long(device=cpu) = onnx::Div(%727, %737) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %739 : Long(device=cpu) = onnx::Cast[to=7](%738) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %740 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%739) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %741 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %742 : Long(device=cpu) = onnx::Div(%730, %741) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %743 : Long(device=cpu) = onnx::Cast[to=7](%742) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %744 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%743) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %745 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %746 : Long(device=cpu) = onnx::Div(%733, %745) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %747 : Long(device=cpu) = onnx::Cast[to=7](%746) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %748 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%747) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %749 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %750 : Long(device=cpu) = onnx::Div(%736, %749) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %751 : Long(device=cpu) = onnx::Cast[to=7](%750) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.19 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%751) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %753 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %754 : Long(requires_grad=0, device=cpu) = onnx::Mul(%740, %753) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %755 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %756 : Long(requires_grad=0, device=cpu) = onnx::Mul(%754, %755) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %757 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %758 : Long(requires_grad=0, device=cpu) = onnx::Mul(%748, %757) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %759 : Long(requires_grad=0, device=cpu) = onnx::Mul(%758, %inferred_length.19) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %761 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %762 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%740, %761)\n", " %765 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %766 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%744, %765)\n", " %767 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %768 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%748, %767)\n", " %769 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %770 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.19, %769)\n", " %771 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%762, %4134, %766, %768, %770) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.51 : Float(*, *, *, *, *, strides=[49152, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%664, %771) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.55 : Float(*, *, *, *, *, strides=[49152, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.51) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %774 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %775 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%756, %774)\n", " %776 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %777 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%759, %776)\n", " %778 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %779 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%744, %778)\n", " %780 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%775, %777, %779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.3 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.55, %780) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %782 : Long(4, strides=[1], device=cpu) = onnx::Shape(%667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %783 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %784 : Long(device=cpu) = onnx::Gather[axis=0](%782, %783) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %785 : Long(4, strides=[1], device=cpu) = onnx::Shape(%667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %786 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %787 : Long(device=cpu) = onnx::Gather[axis=0](%785, %786) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %788 : Long(4, strides=[1], device=cpu) = onnx::Shape(%667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %789 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %790 : Long(device=cpu) = onnx::Gather[axis=0](%788, %789) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %791 : Long(4, strides=[1], device=cpu) = onnx::Shape(%667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %792 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %793 : Long(device=cpu) = onnx::Gather[axis=0](%791, %792) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %794 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %795 : Long(device=cpu) = onnx::Div(%784, %794) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %796 : Long(device=cpu) = onnx::Cast[to=7](%795) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %797 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%796) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %798 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %799 : Long(device=cpu) = onnx::Div(%787, %798) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %800 : Long(device=cpu) = onnx::Cast[to=7](%799) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %801 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%800) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %802 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %803 : Long(device=cpu) = onnx::Div(%790, %802) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %804 : Long(device=cpu) = onnx::Cast[to=7](%803) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %805 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%804) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %806 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %807 : Long(device=cpu) = onnx::Div(%793, %806) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %808 : Long(device=cpu) = onnx::Cast[to=7](%807) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.23 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %810 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %811 : Long(requires_grad=0, device=cpu) = onnx::Mul(%797, %810) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %812 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %813 : Long(requires_grad=0, device=cpu) = onnx::Mul(%811, %812) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %814 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %815 : Long(requires_grad=0, device=cpu) = onnx::Mul(%805, %814) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %816 : Long(requires_grad=0, device=cpu) = onnx::Mul(%815, %inferred_length.23) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %818 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %819 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%797, %818)\n", " %822 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %823 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%801, %822)\n", " %824 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %825 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%805, %824)\n", " %826 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %827 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.23, %826)\n", " %828 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%819, %4135, %823, %825, %827) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.59 : Float(*, *, *, *, *, strides=[49152, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%667, %828) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.63 : Float(*, *, *, *, *, strides=[49152, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.59) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %831 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %832 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%813, %831)\n", " %833 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %834 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%816, %833)\n", " %835 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %836 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%801, %835)\n", " %837 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%832, %834, %836) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.3 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.63, %837) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %839 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.3, %k.3) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %840 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.72 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Mul(%839, %840) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.76 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.72) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.67 : Float(*, *, *, strides=[65536, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.76, %v.3) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %844 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.67) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %845 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %846 : Long(device=cpu) = onnx::Gather[axis=0](%844, %845) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %847 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.67) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %848 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %849 : Long(device=cpu) = onnx::Gather[axis=0](%847, %848) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %850 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.67) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %851 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %852 : Long(device=cpu) = onnx::Gather[axis=0](%850, %851) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %853 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %854 : Long(device=cpu) = onnx::Div(%846, %853) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %855 : Long(device=cpu) = onnx::Cast[to=7](%854) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %856 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%855) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %857 : Long(device=cpu) = onnx::Div(%849, %647) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %858 : Long(device=cpu) = onnx::Cast[to=7](%857) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %859 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%858) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %860 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %861 : Long(device=cpu) = onnx::Div(%852, %860) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %862 : Long(device=cpu) = onnx::Cast[to=7](%861) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.27 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%862) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %864 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]()\n", " %865 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.27, %864) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %867 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %868 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%856, %867)\n", " %871 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %872 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%859, %871)\n", " %873 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %874 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%647, %873)\n", " %875 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %876 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.27, %875)\n", " %877 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%868, %4136, %872, %874, %876) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.71 : Float(*, *, *, *, *, strides=[196608, 65536, 2048, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.67, %877) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.75 : Float(*, *, *, *, *, strides=[196608, 65536, 1, 2048, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.71) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %880 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %881 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%856, %880)\n", " %882 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %883 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%865, %882)\n", " %884 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %885 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%859, %884)\n", " %886 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %887 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%647, %886)\n", " %888 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%881, %883, %885, %887) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.80 : Float(*, *, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.75, %888) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.84 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.80, %layers.1.2.layers.0.0.fn.to_out.0.weight, %layers.1.2.layers.0.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %891 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%input.84, %632) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %892 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%891) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %893 : Float(*, 192, *, *, device=cpu) = onnx::Sub(%891, %892) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %894 : Float(*, 192, *, *, device=cpu) = onnx::Mul(%893, %893) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %895 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%894) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %896 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%891) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %897 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sub(%891, %896) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %898 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %899 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%895, %898) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %900 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%899) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %901 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Div(%897, %900) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %902 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%901, %layers.1.2.layers.0.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.88 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%902, %layers.1.2.layers.0.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %904 : Float(*, 768, *, *, strides=[786432, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.88, %layers.1.2.layers.0.1.fn.net.0.weight, %layers.1.2.layers.0.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %905 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %906 : Float(*, 768, *, *, device=cpu) = onnx::Div(%904, %905) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %907 : Float(*, 768, *, *, device=cpu) = onnx::Erf(%906) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %908 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %909 : Float(*, 768, *, *, device=cpu) = onnx::Add(%907, %908) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %910 : Float(*, 768, *, *, device=cpu) = onnx::Mul(%904, %909) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %911 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.92 : Float(*, 768, *, *, strides=[786432, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%910, %911) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.96 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.92, %layers.1.2.layers.0.1.fn.net.3.weight, %layers.1.2.layers.0.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %914 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%input.96, %891) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %915 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%914) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %916 : Float(*, 192, *, *, device=cpu) = onnx::Sub(%914, %915) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %917 : Float(*, 192, *, *, device=cpu) = onnx::Mul(%916, %916) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %918 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%917) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %919 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%914) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %920 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sub(%914, %919) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %921 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %922 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%918, %921) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %923 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%922) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %924 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Div(%920, %923) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %925 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%924, %layers.1.2.layers.1.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.100 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%925, %layers.1.2.layers.1.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %927 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.100) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %928 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %929 : Long(device=cpu) = onnx::Gather[axis=0](%927, %928) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.108 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=192, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.100, %4064, %4065) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.79 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.108, %layers.1.2.layers.1.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.116 : Float(*, 192, *, *, strides=[49152, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=192, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.100, %4067, %4068) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %935 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.116, %layers.1.2.layers.1.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %936 : Long(4, strides=[1], device=cpu) = onnx::Shape(%935) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %937 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %938 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%936, %937) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %939 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %940 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %941 : Long(1, strides=[1], device=cpu) = onnx::Add(%938, %940) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %942 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %943 : Long(1, strides=[1], device=cpu) = onnx::Div(%941, %942) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %944 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %945 : Long(1, strides=[1], device=cpu) = onnx::Mul(%943, %944) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %946 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Slice(%935, %939, %945, %937) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %947 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %948 : Long(1, strides=[1], device=cpu) = onnx::Mul(%943, %947) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %949 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Slice(%935, %945, %948, %937) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %950 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.79) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %951 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %952 : Long(device=cpu) = onnx::Gather[axis=0](%950, %951) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %953 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.79) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %954 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %955 : Long(device=cpu) = onnx::Gather[axis=0](%953, %954) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %956 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.79) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %957 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %958 : Long(device=cpu) = onnx::Gather[axis=0](%956, %957) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %959 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.79) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %960 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %961 : Long(device=cpu) = onnx::Gather[axis=0](%959, %960) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %962 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %963 : Long(device=cpu) = onnx::Div(%952, %962) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %964 : Long(device=cpu) = onnx::Cast[to=7](%963) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %965 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%964) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %966 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %967 : Long(device=cpu) = onnx::Div(%955, %966) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %968 : Long(device=cpu) = onnx::Cast[to=7](%967) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %969 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%968) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %970 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %971 : Long(device=cpu) = onnx::Div(%958, %970) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %972 : Long(device=cpu) = onnx::Cast[to=7](%971) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %973 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%972) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %974 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %975 : Long(device=cpu) = onnx::Div(%961, %974) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %976 : Long(device=cpu) = onnx::Cast[to=7](%975) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.31 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%976) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %978 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %979 : Long(requires_grad=0, device=cpu) = onnx::Mul(%965, %978) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %980 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %981 : Long(requires_grad=0, device=cpu) = onnx::Mul(%979, %980) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %982 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %983 : Long(requires_grad=0, device=cpu) = onnx::Mul(%973, %982) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %984 : Long(requires_grad=0, device=cpu) = onnx::Mul(%983, %inferred_length.31) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %986 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %987 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%965, %986)\n", " %990 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %991 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%969, %990)\n", " %992 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %993 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%973, %992)\n", " %994 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %995 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.31, %994)\n", " %996 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%987, %4137, %991, %993, %995) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.83 : Float(*, *, *, *, *, strides=[196608, 65536, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.79, %996) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.87 : Float(*, *, *, *, *, strides=[196608, 65536, 32, 1, 1024], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.83) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %999 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1000 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%981, %999)\n", " %1001 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1002 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%984, %1001)\n", " %1003 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1004 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%969, %1003)\n", " %1005 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1000, %1002, %1004) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.7 : Float(*, *, *, strides=[65536, 1, 1024], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.87, %1005) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1007 : Long(4, strides=[1], device=cpu) = onnx::Shape(%946) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1008 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1009 : Long(device=cpu) = onnx::Gather[axis=0](%1007, %1008) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1010 : Long(4, strides=[1], device=cpu) = onnx::Shape(%946) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1011 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1012 : Long(device=cpu) = onnx::Gather[axis=0](%1010, %1011) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1013 : Long(4, strides=[1], device=cpu) = onnx::Shape(%946) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1014 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1015 : Long(device=cpu) = onnx::Gather[axis=0](%1013, %1014) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1016 : Long(4, strides=[1], device=cpu) = onnx::Shape(%946) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1017 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1018 : Long(device=cpu) = onnx::Gather[axis=0](%1016, %1017) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1019 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1020 : Long(device=cpu) = onnx::Div(%1009, %1019) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1021 : Long(device=cpu) = onnx::Cast[to=7](%1020) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1022 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1021) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1023 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1024 : Long(device=cpu) = onnx::Div(%1012, %1023) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1025 : Long(device=cpu) = onnx::Cast[to=7](%1024) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1026 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1025) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1027 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1028 : Long(device=cpu) = onnx::Div(%1015, %1027) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1029 : Long(device=cpu) = onnx::Cast[to=7](%1028) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1030 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1029) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1031 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1032 : Long(device=cpu) = onnx::Div(%1018, %1031) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1033 : Long(device=cpu) = onnx::Cast[to=7](%1032) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.35 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1033) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1035 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1036 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1022, %1035) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1037 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1038 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1036, %1037) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1039 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1040 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1030, %1039) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1041 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1040, %inferred_length.35) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1043 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1044 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1022, %1043)\n", " %1047 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1048 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1026, %1047)\n", " %1049 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1050 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1030, %1049)\n", " %1051 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1052 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.35, %1051)\n", " %1053 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1044, %4138, %1048, %1050, %1052) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.91 : Float(*, *, *, *, *, strides=[49152, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%946, %1053) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.95 : Float(*, *, *, *, *, strides=[49152, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.91) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1056 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1057 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1038, %1056)\n", " %1058 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1059 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1041, %1058)\n", " %1060 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1061 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1026, %1060)\n", " %1062 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1057, %1059, %1061) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.7 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.95, %1062) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1064 : Long(4, strides=[1], device=cpu) = onnx::Shape(%949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1065 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1066 : Long(device=cpu) = onnx::Gather[axis=0](%1064, %1065) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1067 : Long(4, strides=[1], device=cpu) = onnx::Shape(%949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1068 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1069 : Long(device=cpu) = onnx::Gather[axis=0](%1067, %1068) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1070 : Long(4, strides=[1], device=cpu) = onnx::Shape(%949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1071 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1072 : Long(device=cpu) = onnx::Gather[axis=0](%1070, %1071) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1073 : Long(4, strides=[1], device=cpu) = onnx::Shape(%949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1074 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1075 : Long(device=cpu) = onnx::Gather[axis=0](%1073, %1074) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1076 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1077 : Long(device=cpu) = onnx::Div(%1066, %1076) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1078 : Long(device=cpu) = onnx::Cast[to=7](%1077) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1079 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1078) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1080 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1081 : Long(device=cpu) = onnx::Div(%1069, %1080) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1082 : Long(device=cpu) = onnx::Cast[to=7](%1081) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1083 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1082) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1084 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1085 : Long(device=cpu) = onnx::Div(%1072, %1084) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1086 : Long(device=cpu) = onnx::Cast[to=7](%1085) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1087 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1086) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1088 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1089 : Long(device=cpu) = onnx::Div(%1075, %1088) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1090 : Long(device=cpu) = onnx::Cast[to=7](%1089) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.39 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1090) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1092 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1093 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1079, %1092) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1094 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1095 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1093, %1094) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1096 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1097 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1087, %1096) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1098 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1097, %inferred_length.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1100 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1101 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1079, %1100)\n", " %1104 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1105 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1083, %1104)\n", " %1106 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1107 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1087, %1106)\n", " %1108 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1109 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.39, %1108)\n", " %1110 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1101, %4139, %1105, %1107, %1109) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.99 : Float(*, *, *, *, *, strides=[49152, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%949, %1110) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.103 : Float(*, *, *, *, *, strides=[49152, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.99) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1113 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1114 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1095, %1113)\n", " %1115 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1116 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1098, %1115)\n", " %1117 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1118 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1083, %1117)\n", " %1119 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1114, %1116, %1118) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.7 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.103, %1119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1121 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.7, %k.7) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1122 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.120 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Mul(%1121, %1122) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.124 : Float(*, *, *, strides=[262144, 256, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.120) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.107 : Float(*, *, *, strides=[65536, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.124, %v.7) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1126 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.107) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1127 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1128 : Long(device=cpu) = onnx::Gather[axis=0](%1126, %1127) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1129 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.107) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1130 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1131 : Long(device=cpu) = onnx::Gather[axis=0](%1129, %1130) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1132 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.107) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1133 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1134 : Long(device=cpu) = onnx::Gather[axis=0](%1132, %1133) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1135 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1136 : Long(device=cpu) = onnx::Div(%1128, %1135) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1137 : Long(device=cpu) = onnx::Cast[to=7](%1136) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1138 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1137) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1139 : Long(device=cpu) = onnx::Div(%1131, %929) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1140 : Long(device=cpu) = onnx::Cast[to=7](%1139) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1141 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1140) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1142 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1143 : Long(device=cpu) = onnx::Div(%1134, %1142) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1144 : Long(device=cpu) = onnx::Cast[to=7](%1143) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.43 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1144) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1146 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={3}]()\n", " %1147 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.43, %1146) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1149 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1150 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1138, %1149)\n", " %1153 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1154 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1141, %1153)\n", " %1155 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1156 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%929, %1155)\n", " %1157 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1158 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.43, %1157)\n", " %1159 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1150, %4140, %1154, %1156, %1158) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.111 : Float(*, *, *, *, *, strides=[196608, 65536, 2048, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.107, %1159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.115 : Float(*, *, *, *, *, strides=[196608, 65536, 1, 2048, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.111) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1162 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1163 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1138, %1162)\n", " %1164 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1165 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1147, %1164)\n", " %1166 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1167 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1141, %1166)\n", " %1168 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1169 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%929, %1168)\n", " %1170 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%1163, %1165, %1167, %1169) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.128 : Float(*, *, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.115, %1170) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.132 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.128, %layers.1.2.layers.1.0.fn.to_out.0.weight, %layers.1.2.layers.1.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1173 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%input.132, %914) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %1174 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1173) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1175 : Float(*, 192, *, *, device=cpu) = onnx::Sub(%1173, %1174) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1176 : Float(*, 192, *, *, device=cpu) = onnx::Mul(%1175, %1175) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1177 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1176) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1178 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1173) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1179 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sub(%1173, %1178) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1180 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1181 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%1177, %1180) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1182 : Float(*, 1, *, *, strides=[1024, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1181) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1183 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Div(%1179, %1182) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1184 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%1183, %layers.1.2.layers.1.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.136 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%1184, %layers.1.2.layers.1.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1186 : Float(*, 768, *, *, strides=[786432, 1024, 32, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.136, %layers.1.2.layers.1.1.fn.net.0.weight, %layers.1.2.layers.1.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1187 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1188 : Float(*, 768, *, *, device=cpu) = onnx::Div(%1186, %1187) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1189 : Float(*, 768, *, *, device=cpu) = onnx::Erf(%1188) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1190 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1191 : Float(*, 768, *, *, device=cpu) = onnx::Add(%1189, %1190) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1192 : Float(*, 768, *, *, device=cpu) = onnx::Mul(%1186, %1191) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1193 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.140 : Float(*, 768, *, *, strides=[786432, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Mul(%1192, %1193) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.144 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.140, %layers.1.2.layers.1.1.fn.net.3.weight, %layers.1.2.layers.1.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.148 : Float(*, 192, *, *, strides=[196608, 1024, 32, 1], requires_grad=1, device=cpu) = onnx::Add(%input.144, %1173) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %1197 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.148, %layers.2.0.weight, %layers.2.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1198 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1197) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1199 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1197, %1198) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1200 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1199, %1199) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1201 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1200) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1202 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1197) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1203 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1197, %1202) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1204 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1205 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1201, %1204) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1206 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1205) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1207 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1203, %1206) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1208 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1207, %layers.2.1.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1209 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1208, %layers.2.1.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1210 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1209) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1211 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1209, %1210) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1212 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1211, %1211) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1213 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1212) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1214 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1209) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1215 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1209, %1214) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1216 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1217 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1213, %1216) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1218 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1217) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1219 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1215, %1218) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1220 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1219, %layers.2.2.layers.0.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.152 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1220, %layers.2.2.layers.0.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1222 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.152) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1223 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1224 : Long(device=cpu) = onnx::Gather[axis=0](%1222, %1223) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.160 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.152, %4070, %4071) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.119 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.160, %layers.2.2.layers.0.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.168 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.152, %4073, %4074) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1230 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.168, %layers.2.2.layers.0.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1231 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1230) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1232 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1233 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%1231, %1232) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1234 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1235 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1236 : Long(1, strides=[1], device=cpu) = onnx::Add(%1233, %1235) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1237 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1238 : Long(1, strides=[1], device=cpu) = onnx::Div(%1236, %1237) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1239 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1240 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1238, %1239) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1241 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1230, %1234, %1240, %1232) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1242 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1243 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1238, %1242) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1244 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1230, %1240, %1243, %1232) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1245 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1246 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1247 : Long(device=cpu) = onnx::Gather[axis=0](%1245, %1246) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1248 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1249 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1250 : Long(device=cpu) = onnx::Gather[axis=0](%1248, %1249) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1251 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1252 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1253 : Long(device=cpu) = onnx::Gather[axis=0](%1251, %1252) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1254 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1255 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1256 : Long(device=cpu) = onnx::Gather[axis=0](%1254, %1255) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1257 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1258 : Long(device=cpu) = onnx::Div(%1247, %1257) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1259 : Long(device=cpu) = onnx::Cast[to=7](%1258) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1260 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1259) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1261 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1262 : Long(device=cpu) = onnx::Div(%1250, %1261) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1263 : Long(device=cpu) = onnx::Cast[to=7](%1262) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1264 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1263) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1265 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1266 : Long(device=cpu) = onnx::Div(%1253, %1265) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1267 : Long(device=cpu) = onnx::Cast[to=7](%1266) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1268 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1267) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1269 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1270 : Long(device=cpu) = onnx::Div(%1256, %1269) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1271 : Long(device=cpu) = onnx::Cast[to=7](%1270) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.47 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1271) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1273 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1274 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1260, %1273) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1275 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1276 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1274, %1275) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1277 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1278 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1268, %1277) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1279 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1278, %inferred_length.47) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1281 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1282 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1260, %1281)\n", " %1285 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1286 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1264, %1285)\n", " %1287 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1288 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1268, %1287)\n", " %1289 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1290 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.47, %1289)\n", " %1291 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1282, %4141, %1286, %1288, %1290) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.123 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.119, %1291) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.127 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.123) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1294 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1295 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1276, %1294)\n", " %1296 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1297 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1279, %1296)\n", " %1298 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1299 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1264, %1298)\n", " %1300 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1295, %1297, %1299) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.11 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.127, %1300) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1302 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1241) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1303 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1304 : Long(device=cpu) = onnx::Gather[axis=0](%1302, %1303) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1305 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1241) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1306 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1307 : Long(device=cpu) = onnx::Gather[axis=0](%1305, %1306) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1308 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1241) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1309 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1310 : Long(device=cpu) = onnx::Gather[axis=0](%1308, %1309) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1311 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1241) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1312 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1313 : Long(device=cpu) = onnx::Gather[axis=0](%1311, %1312) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1314 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1315 : Long(device=cpu) = onnx::Div(%1304, %1314) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1316 : Long(device=cpu) = onnx::Cast[to=7](%1315) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1317 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1316) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1318 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1319 : Long(device=cpu) = onnx::Div(%1307, %1318) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1320 : Long(device=cpu) = onnx::Cast[to=7](%1319) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1321 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1320) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1322 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1323 : Long(device=cpu) = onnx::Div(%1310, %1322) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1324 : Long(device=cpu) = onnx::Cast[to=7](%1323) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1325 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1324) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1326 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1327 : Long(device=cpu) = onnx::Div(%1313, %1326) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1328 : Long(device=cpu) = onnx::Cast[to=7](%1327) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.51 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1328) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1330 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1331 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1317, %1330) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1332 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1333 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1331, %1332) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1334 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1335 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1325, %1334) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1336 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1335, %inferred_length.51) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1338 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1339 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1317, %1338)\n", " %1342 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1343 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1321, %1342)\n", " %1344 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1345 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1325, %1344)\n", " %1346 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1347 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.51, %1346)\n", " %1348 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1339, %4142, %1343, %1345, %1347) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.131 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1241, %1348) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.135 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.131) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1351 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1352 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1333, %1351)\n", " %1353 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1354 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1336, %1353)\n", " %1355 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1356 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1321, %1355)\n", " %1357 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1352, %1354, %1356) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.11 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.135, %1357) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1359 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1244) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1360 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1361 : Long(device=cpu) = onnx::Gather[axis=0](%1359, %1360) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1362 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1244) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1363 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1364 : Long(device=cpu) = onnx::Gather[axis=0](%1362, %1363) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1365 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1244) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1366 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1367 : Long(device=cpu) = onnx::Gather[axis=0](%1365, %1366) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1368 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1244) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1369 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1370 : Long(device=cpu) = onnx::Gather[axis=0](%1368, %1369) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1371 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1372 : Long(device=cpu) = onnx::Div(%1361, %1371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1373 : Long(device=cpu) = onnx::Cast[to=7](%1372) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1374 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1373) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1375 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1376 : Long(device=cpu) = onnx::Div(%1364, %1375) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1377 : Long(device=cpu) = onnx::Cast[to=7](%1376) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1378 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1377) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1379 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1380 : Long(device=cpu) = onnx::Div(%1367, %1379) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1381 : Long(device=cpu) = onnx::Cast[to=7](%1380) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1382 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1381) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1383 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1384 : Long(device=cpu) = onnx::Div(%1370, %1383) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1385 : Long(device=cpu) = onnx::Cast[to=7](%1384) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.55 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1385) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1387 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1388 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1374, %1387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1389 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1390 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1388, %1389) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1391 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1392 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1382, %1391) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1393 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1392, %inferred_length.55) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1395 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1396 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1374, %1395)\n", " %1399 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1400 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1378, %1399)\n", " %1401 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1402 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1382, %1401)\n", " %1403 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1404 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.55, %1403)\n", " %1405 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1396, %4143, %1400, %1402, %1404) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.139 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1244, %1405) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.143 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.139) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1408 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1409 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1390, %1408)\n", " %1410 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1411 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1393, %1410)\n", " %1412 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1413 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1378, %1412)\n", " %1414 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1409, %1411, %1413) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.11 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.143, %1414) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1416 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.11, %k.11) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1417 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.172 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%1416, %1417) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.176 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.172) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.147 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.176, %v.11) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1421 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.147) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1422 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1423 : Long(device=cpu) = onnx::Gather[axis=0](%1421, %1422) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1424 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.147) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1425 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1426 : Long(device=cpu) = onnx::Gather[axis=0](%1424, %1425) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1427 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.147) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1428 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1429 : Long(device=cpu) = onnx::Gather[axis=0](%1427, %1428) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1430 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1431 : Long(device=cpu) = onnx::Div(%1423, %1430) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1432 : Long(device=cpu) = onnx::Cast[to=7](%1431) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1433 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1432) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1434 : Long(device=cpu) = onnx::Div(%1426, %1224) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1435 : Long(device=cpu) = onnx::Cast[to=7](%1434) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1436 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1435) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1437 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1438 : Long(device=cpu) = onnx::Div(%1429, %1437) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1439 : Long(device=cpu) = onnx::Cast[to=7](%1438) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.59 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1441 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %1442 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.59, %1441) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1444 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1445 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1433, %1444)\n", " %1448 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1449 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1436, %1448)\n", " %1450 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1451 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1224, %1450)\n", " %1452 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1453 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.59, %1452)\n", " %1454 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1445, %4144, %1449, %1451, %1453) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.151 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.147, %1454) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.155 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.151) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1457 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1458 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1433, %1457)\n", " %1459 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1460 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1442, %1459)\n", " %1461 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1462 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1436, %1461)\n", " %1463 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1464 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1224, %1463)\n", " %1465 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%1458, %1460, %1462, %1464) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.180 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.155, %1465) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.184 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.180, %layers.2.2.layers.0.0.fn.to_out.0.weight, %layers.2.2.layers.0.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1468 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.184, %1209) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %1469 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1468) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1470 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1468, %1469) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1471 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1470, %1470) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1472 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1471) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1473 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1468) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1474 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1468, %1473) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1475 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1476 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1472, %1475) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1477 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1476) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1478 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1474, %1477) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1479 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1478, %layers.2.2.layers.0.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.188 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1479, %layers.2.2.layers.0.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1481 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.188, %layers.2.2.layers.0.1.fn.net.0.weight, %layers.2.2.layers.0.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1482 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1483 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%1481, %1482) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1484 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%1483) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1485 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1486 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%1484, %1485) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1487 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%1481, %1486) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1488 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.192 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1487, %1488) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.196 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.192, %layers.2.2.layers.0.1.fn.net.3.weight, %layers.2.2.layers.0.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1491 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.196, %1468) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %1492 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1491) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1493 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1491, %1492) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1494 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1493, %1493) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1495 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1494) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1496 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1491) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1497 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1491, %1496) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1498 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1499 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1495, %1498) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1500 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1499) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1501 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1497, %1500) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1502 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1501, %layers.2.2.layers.1.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.200 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1502, %layers.2.2.layers.1.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1504 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.200) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1505 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1506 : Long(device=cpu) = onnx::Gather[axis=0](%1504, %1505) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.208 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.200, %4076, %4077) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.159 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.208, %layers.2.2.layers.1.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.216 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.200, %4079, %4080) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1512 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.216, %layers.2.2.layers.1.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1513 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1512) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1514 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1515 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%1513, %1514) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1516 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1517 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1518 : Long(1, strides=[1], device=cpu) = onnx::Add(%1515, %1517) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1519 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1520 : Long(1, strides=[1], device=cpu) = onnx::Div(%1518, %1519) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1521 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1522 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1520, %1521) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1523 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1512, %1516, %1522, %1514) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1524 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1525 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1520, %1524) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1526 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1512, %1522, %1525, %1514) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1527 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1528 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1529 : Long(device=cpu) = onnx::Gather[axis=0](%1527, %1528) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1530 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1531 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1532 : Long(device=cpu) = onnx::Gather[axis=0](%1530, %1531) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1533 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1534 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1535 : Long(device=cpu) = onnx::Gather[axis=0](%1533, %1534) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1536 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1537 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1538 : Long(device=cpu) = onnx::Gather[axis=0](%1536, %1537) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1539 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1540 : Long(device=cpu) = onnx::Div(%1529, %1539) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1541 : Long(device=cpu) = onnx::Cast[to=7](%1540) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1542 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1541) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1543 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1544 : Long(device=cpu) = onnx::Div(%1532, %1543) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1545 : Long(device=cpu) = onnx::Cast[to=7](%1544) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1546 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1545) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1547 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1548 : Long(device=cpu) = onnx::Div(%1535, %1547) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1549 : Long(device=cpu) = onnx::Cast[to=7](%1548) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1550 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1549) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1551 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1552 : Long(device=cpu) = onnx::Div(%1538, %1551) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1553 : Long(device=cpu) = onnx::Cast[to=7](%1552) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.63 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1553) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1555 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1556 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1542, %1555) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1557 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1558 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1556, %1557) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1559 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1560 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1550, %1559) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1561 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1560, %inferred_length.63) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1563 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1564 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1542, %1563)\n", " %1567 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1568 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1546, %1567)\n", " %1569 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1570 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1550, %1569)\n", " %1571 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1572 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.63, %1571)\n", " %1573 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1564, %4145, %1568, %1570, %1572) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.163 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.159, %1573) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.167 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.163) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1576 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1577 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1558, %1576)\n", " %1578 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1579 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1561, %1578)\n", " %1580 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1581 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1546, %1580)\n", " %1582 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1577, %1579, %1581) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.15 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.167, %1582) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1584 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1523) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1585 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1586 : Long(device=cpu) = onnx::Gather[axis=0](%1584, %1585) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1587 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1523) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1588 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1589 : Long(device=cpu) = onnx::Gather[axis=0](%1587, %1588) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1590 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1523) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1591 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1592 : Long(device=cpu) = onnx::Gather[axis=0](%1590, %1591) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1593 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1523) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1594 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1595 : Long(device=cpu) = onnx::Gather[axis=0](%1593, %1594) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1596 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1597 : Long(device=cpu) = onnx::Div(%1586, %1596) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1598 : Long(device=cpu) = onnx::Cast[to=7](%1597) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1599 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1598) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1600 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1601 : Long(device=cpu) = onnx::Div(%1589, %1600) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1602 : Long(device=cpu) = onnx::Cast[to=7](%1601) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1603 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1602) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1604 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1605 : Long(device=cpu) = onnx::Div(%1592, %1604) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1606 : Long(device=cpu) = onnx::Cast[to=7](%1605) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1607 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1606) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1608 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1609 : Long(device=cpu) = onnx::Div(%1595, %1608) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1610 : Long(device=cpu) = onnx::Cast[to=7](%1609) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.67 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1610) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1612 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1613 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1599, %1612) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1614 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1615 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1613, %1614) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1616 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1617 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1607, %1616) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1618 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1617, %inferred_length.67) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1620 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1621 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1599, %1620)\n", " %1624 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1625 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1603, %1624)\n", " %1626 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1627 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1607, %1626)\n", " %1628 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1629 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.67, %1628)\n", " %1630 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1621, %4146, %1625, %1627, %1629) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.171 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1523, %1630) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.175 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.171) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1633 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1634 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1615, %1633)\n", " %1635 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1636 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1618, %1635)\n", " %1637 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1638 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1603, %1637)\n", " %1639 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1634, %1636, %1638) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.15 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.175, %1639) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1641 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1526) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1642 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1643 : Long(device=cpu) = onnx::Gather[axis=0](%1641, %1642) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1644 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1526) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1645 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1646 : Long(device=cpu) = onnx::Gather[axis=0](%1644, %1645) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1647 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1526) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1648 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1649 : Long(device=cpu) = onnx::Gather[axis=0](%1647, %1648) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1650 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1526) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1651 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1652 : Long(device=cpu) = onnx::Gather[axis=0](%1650, %1651) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1653 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1654 : Long(device=cpu) = onnx::Div(%1643, %1653) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1655 : Long(device=cpu) = onnx::Cast[to=7](%1654) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1656 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1655) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1657 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1658 : Long(device=cpu) = onnx::Div(%1646, %1657) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1659 : Long(device=cpu) = onnx::Cast[to=7](%1658) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1660 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1659) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1661 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1662 : Long(device=cpu) = onnx::Div(%1649, %1661) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1663 : Long(device=cpu) = onnx::Cast[to=7](%1662) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1664 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1663) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1665 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1666 : Long(device=cpu) = onnx::Div(%1652, %1665) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1667 : Long(device=cpu) = onnx::Cast[to=7](%1666) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.71 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1669 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1670 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1656, %1669) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1671 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1672 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1670, %1671) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1673 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1674 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1664, %1673) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1675 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1674, %inferred_length.71) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1677 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1678 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1656, %1677)\n", " %1681 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1682 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1660, %1681)\n", " %1683 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1684 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1664, %1683)\n", " %1685 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1686 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.71, %1685)\n", " %1687 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1678, %4147, %1682, %1684, %1686) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.179 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1526, %1687) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.183 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.179) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1690 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1691 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1672, %1690)\n", " %1692 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1693 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1675, %1692)\n", " %1694 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1695 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1660, %1694)\n", " %1696 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1691, %1693, %1695) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.15 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.183, %1696) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1698 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.15, %k.15) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1699 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.220 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%1698, %1699) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.224 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.220) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.187 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.224, %v.15) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1703 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.187) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1704 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1705 : Long(device=cpu) = onnx::Gather[axis=0](%1703, %1704) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1706 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.187) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1707 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1708 : Long(device=cpu) = onnx::Gather[axis=0](%1706, %1707) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1709 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.187) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1710 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1711 : Long(device=cpu) = onnx::Gather[axis=0](%1709, %1710) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1712 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1713 : Long(device=cpu) = onnx::Div(%1705, %1712) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1714 : Long(device=cpu) = onnx::Cast[to=7](%1713) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1715 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1714) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1716 : Long(device=cpu) = onnx::Div(%1708, %1506) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1717 : Long(device=cpu) = onnx::Cast[to=7](%1716) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1718 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1717) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1719 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1720 : Long(device=cpu) = onnx::Div(%1711, %1719) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1721 : Long(device=cpu) = onnx::Cast[to=7](%1720) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.75 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1721) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1723 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %1724 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.75, %1723) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1726 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1727 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1715, %1726)\n", " %1730 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1731 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1718, %1730)\n", " %1732 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1733 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1506, %1732)\n", " %1734 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1735 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.75, %1734)\n", " %1736 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1727, %4148, %1731, %1733, %1735) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.191 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.187, %1736) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.195 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.191) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1739 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1740 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1715, %1739)\n", " %1741 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1742 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1724, %1741)\n", " %1743 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1744 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1718, %1743)\n", " %1745 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1746 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1506, %1745)\n", " %1747 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%1740, %1742, %1744, %1746) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.228 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.195, %1747) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.232 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.228, %layers.2.2.layers.1.0.fn.to_out.0.weight, %layers.2.2.layers.1.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1750 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.232, %1491) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %1751 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1750) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1752 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1750, %1751) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1753 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1752, %1752) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1754 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1753) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1755 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1750) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1756 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1750, %1755) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1757 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1758 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1754, %1757) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1759 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1758) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1760 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1756, %1759) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1761 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1760, %layers.2.2.layers.1.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.236 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1761, %layers.2.2.layers.1.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1763 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.236, %layers.2.2.layers.1.1.fn.net.0.weight, %layers.2.2.layers.1.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1764 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1765 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%1763, %1764) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1766 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%1765) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1767 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1768 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%1766, %1767) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1769 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%1763, %1768) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %1770 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.240 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1769, %1770) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.244 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.240, %layers.2.2.layers.1.1.fn.net.3.weight, %layers.2.2.layers.1.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1773 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.244, %1750) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %1774 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1773) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1775 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%1773, %1774) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1776 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%1775, %1775) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1777 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1776) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %1778 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%1773) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %1779 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%1773, %1778) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1780 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %1781 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1777, %1780) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1782 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%1781) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1783 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%1779, %1782) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1784 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%1783, %layers.2.2.layers.2.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.248 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%1784, %layers.2.2.layers.2.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %1786 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.248) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1787 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %1788 : Long(device=cpu) = onnx::Gather[axis=0](%1786, %1787) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.256 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.248, %4082, %4083) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.199 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.256, %layers.2.2.layers.2.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.264 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.248, %4085, %4086) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1794 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.264, %layers.2.2.layers.2.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %1795 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1794) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1796 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1797 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%1795, %1796) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1798 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1799 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1800 : Long(1, strides=[1], device=cpu) = onnx::Add(%1797, %1799) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1801 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1802 : Long(1, strides=[1], device=cpu) = onnx::Div(%1800, %1801) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1803 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1804 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1802, %1803) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1805 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1794, %1798, %1804, %1796) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1806 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1807 : Long(1, strides=[1], device=cpu) = onnx::Mul(%1802, %1806) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1808 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%1794, %1804, %1807, %1796) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %1809 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.199) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1810 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1811 : Long(device=cpu) = onnx::Gather[axis=0](%1809, %1810) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1812 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.199) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1813 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1814 : Long(device=cpu) = onnx::Gather[axis=0](%1812, %1813) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1815 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.199) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1816 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1817 : Long(device=cpu) = onnx::Gather[axis=0](%1815, %1816) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1818 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.199) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1819 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1820 : Long(device=cpu) = onnx::Gather[axis=0](%1818, %1819) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1821 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1822 : Long(device=cpu) = onnx::Div(%1811, %1821) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1823 : Long(device=cpu) = onnx::Cast[to=7](%1822) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1824 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1823) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1825 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1826 : Long(device=cpu) = onnx::Div(%1814, %1825) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1827 : Long(device=cpu) = onnx::Cast[to=7](%1826) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1828 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1827) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1829 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1830 : Long(device=cpu) = onnx::Div(%1817, %1829) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1831 : Long(device=cpu) = onnx::Cast[to=7](%1830) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1832 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1831) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1833 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1834 : Long(device=cpu) = onnx::Div(%1820, %1833) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1835 : Long(device=cpu) = onnx::Cast[to=7](%1834) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.79 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1835) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1837 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1838 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1824, %1837) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1839 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1840 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1838, %1839) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1841 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1842 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1832, %1841) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1843 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1842, %inferred_length.79) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1845 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1846 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1824, %1845)\n", " %1849 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1850 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1828, %1849)\n", " %1851 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1852 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1832, %1851)\n", " %1853 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1854 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.79, %1853)\n", " %1855 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1846, %4149, %1850, %1852, %1854) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.203 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.199, %1855) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.207 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.203) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1858 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1859 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1840, %1858)\n", " %1860 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1861 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1843, %1860)\n", " %1862 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1863 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1828, %1862)\n", " %1864 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1859, %1861, %1863) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.19 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.207, %1864) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1866 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1805) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1867 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1868 : Long(device=cpu) = onnx::Gather[axis=0](%1866, %1867) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1869 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1805) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1870 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1871 : Long(device=cpu) = onnx::Gather[axis=0](%1869, %1870) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1872 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1805) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1873 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1874 : Long(device=cpu) = onnx::Gather[axis=0](%1872, %1873) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1875 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1805) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1876 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1877 : Long(device=cpu) = onnx::Gather[axis=0](%1875, %1876) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1878 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1879 : Long(device=cpu) = onnx::Div(%1868, %1878) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1880 : Long(device=cpu) = onnx::Cast[to=7](%1879) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1881 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1880) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1882 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1883 : Long(device=cpu) = onnx::Div(%1871, %1882) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1884 : Long(device=cpu) = onnx::Cast[to=7](%1883) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1885 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1884) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1886 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1887 : Long(device=cpu) = onnx::Div(%1874, %1886) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1888 : Long(device=cpu) = onnx::Cast[to=7](%1887) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1889 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1888) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1890 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1891 : Long(device=cpu) = onnx::Div(%1877, %1890) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1892 : Long(device=cpu) = onnx::Cast[to=7](%1891) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.83 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1892) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1894 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1895 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1881, %1894) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1896 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1897 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1895, %1896) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1898 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1899 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1889, %1898) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1900 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1899, %inferred_length.83) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1902 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1903 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1881, %1902)\n", " %1906 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1907 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1885, %1906)\n", " %1908 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1909 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1889, %1908)\n", " %1910 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1911 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.83, %1910)\n", " %1912 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1903, %4150, %1907, %1909, %1911) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.211 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1805, %1912) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.215 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.211) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1915 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1916 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1897, %1915)\n", " %1917 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1918 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1900, %1917)\n", " %1919 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1920 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1885, %1919)\n", " %1921 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1916, %1918, %1920) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.19 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.215, %1921) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1923 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1924 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1925 : Long(device=cpu) = onnx::Gather[axis=0](%1923, %1924) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1926 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1927 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1928 : Long(device=cpu) = onnx::Gather[axis=0](%1926, %1927) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1929 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1930 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1931 : Long(device=cpu) = onnx::Gather[axis=0](%1929, %1930) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1932 : Long(4, strides=[1], device=cpu) = onnx::Shape(%1808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1933 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1934 : Long(device=cpu) = onnx::Gather[axis=0](%1932, %1933) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1935 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1936 : Long(device=cpu) = onnx::Div(%1925, %1935) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1937 : Long(device=cpu) = onnx::Cast[to=7](%1936) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1938 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1937) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1939 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1940 : Long(device=cpu) = onnx::Div(%1928, %1939) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1941 : Long(device=cpu) = onnx::Cast[to=7](%1940) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1942 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1941) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1943 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1944 : Long(device=cpu) = onnx::Div(%1931, %1943) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1945 : Long(device=cpu) = onnx::Cast[to=7](%1944) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1946 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1945) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1947 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1948 : Long(device=cpu) = onnx::Div(%1934, %1947) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1949 : Long(device=cpu) = onnx::Cast[to=7](%1948) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.87 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1951 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1952 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1938, %1951) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1953 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1954 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1952, %1953) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1955 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %1956 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1946, %1955) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1957 : Long(requires_grad=0, device=cpu) = onnx::Mul(%1956, %inferred_length.87) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %1959 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1960 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1938, %1959)\n", " %1963 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1964 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1942, %1963)\n", " %1965 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1966 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1946, %1965)\n", " %1967 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1968 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.87, %1967)\n", " %1969 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%1960, %4151, %1964, %1966, %1968) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.219 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%1808, %1969) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.223 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.219) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %1972 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1973 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1954, %1972)\n", " %1974 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1975 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1957, %1974)\n", " %1976 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %1977 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1942, %1976)\n", " %1978 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%1973, %1975, %1977) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.19 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.223, %1978) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %1980 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.19, %k.19) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1981 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.268 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%1980, %1981) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.272 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.268) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.227 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.272, %v.19) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %1985 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.227) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1986 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1987 : Long(device=cpu) = onnx::Gather[axis=0](%1985, %1986) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1988 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.227) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1989 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1990 : Long(device=cpu) = onnx::Gather[axis=0](%1988, %1989) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1991 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.227) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1992 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1993 : Long(device=cpu) = onnx::Gather[axis=0](%1991, %1992) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %1994 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1995 : Long(device=cpu) = onnx::Div(%1987, %1994) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1996 : Long(device=cpu) = onnx::Cast[to=7](%1995) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1997 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1996) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1998 : Long(device=cpu) = onnx::Div(%1990, %1788) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %1999 : Long(device=cpu) = onnx::Cast[to=7](%1998) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2000 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%1999) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2001 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2002 : Long(device=cpu) = onnx::Div(%1993, %2001) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2003 : Long(device=cpu) = onnx::Cast[to=7](%2002) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.91 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2003) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2005 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %2006 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.91, %2005) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2008 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2009 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1997, %2008)\n", " %2012 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2013 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2000, %2012)\n", " %2014 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2015 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1788, %2014)\n", " %2016 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2017 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.91, %2016)\n", " %2018 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2009, %4152, %2013, %2015, %2017) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.231 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.227, %2018) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.235 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.231) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2021 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2022 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1997, %2021)\n", " %2023 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2024 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2006, %2023)\n", " %2025 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2026 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2000, %2025)\n", " %2027 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2028 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%1788, %2027)\n", " %2029 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%2022, %2024, %2026, %2028) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.276 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.235, %2029) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.280 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.276, %layers.2.2.layers.2.0.fn.to_out.0.weight, %layers.2.2.layers.2.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2032 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.280, %1773) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %2033 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2032) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2034 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2032, %2033) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2035 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2034, %2034) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2036 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2035) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2037 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2032) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2038 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2032, %2037) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2039 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2040 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2036, %2039) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2041 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2040) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2042 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2038, %2041) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2043 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2042, %layers.2.2.layers.2.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.284 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2043, %layers.2.2.layers.2.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2045 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.284, %layers.2.2.layers.2.1.fn.net.0.weight, %layers.2.2.layers.2.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2046 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2047 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%2045, %2046) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2048 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%2047) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2049 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2050 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%2048, %2049) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2051 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%2045, %2050) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2052 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.288 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2051, %2052) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.292 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.288, %layers.2.2.layers.2.1.fn.net.3.weight, %layers.2.2.layers.2.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2055 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.292, %2032) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %2056 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2055) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2057 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2055, %2056) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2058 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2057, %2057) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2059 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2058) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2060 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2055) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2061 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2055, %2060) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2062 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2063 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2059, %2062) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2064 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2063) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2065 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2061, %2064) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2066 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2065, %layers.2.2.layers.3.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.296 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2066, %layers.2.2.layers.3.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2068 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.296) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2069 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2070 : Long(device=cpu) = onnx::Gather[axis=0](%2068, %2069) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.304 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.296, %4088, %4089) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.239 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.304, %layers.2.2.layers.3.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.312 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.296, %4091, %4092) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2076 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.312, %layers.2.2.layers.3.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2077 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2076) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2078 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2079 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%2077, %2078) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2080 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2081 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2082 : Long(1, strides=[1], device=cpu) = onnx::Add(%2079, %2081) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2083 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2084 : Long(1, strides=[1], device=cpu) = onnx::Div(%2082, %2083) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2085 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2086 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2084, %2085) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2087 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2076, %2080, %2086, %2078) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2088 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2089 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2084, %2088) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2090 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2076, %2086, %2089, %2078) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2091 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.239) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2092 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2093 : Long(device=cpu) = onnx::Gather[axis=0](%2091, %2092) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2094 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.239) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2095 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2096 : Long(device=cpu) = onnx::Gather[axis=0](%2094, %2095) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2097 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.239) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2098 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2099 : Long(device=cpu) = onnx::Gather[axis=0](%2097, %2098) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2100 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.239) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2101 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2102 : Long(device=cpu) = onnx::Gather[axis=0](%2100, %2101) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2103 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2104 : Long(device=cpu) = onnx::Div(%2093, %2103) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2105 : Long(device=cpu) = onnx::Cast[to=7](%2104) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2106 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2105) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2107 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2108 : Long(device=cpu) = onnx::Div(%2096, %2107) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2109 : Long(device=cpu) = onnx::Cast[to=7](%2108) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2110 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2109) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2111 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2112 : Long(device=cpu) = onnx::Div(%2099, %2111) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2113 : Long(device=cpu) = onnx::Cast[to=7](%2112) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2114 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2113) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2115 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2116 : Long(device=cpu) = onnx::Div(%2102, %2115) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2117 : Long(device=cpu) = onnx::Cast[to=7](%2116) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.95 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2117) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2119 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2120 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2106, %2119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2121 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2122 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2120, %2121) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2123 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2124 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2114, %2123) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2125 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2124, %inferred_length.95) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2127 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2128 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2106, %2127)\n", " %2131 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2132 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2110, %2131)\n", " %2133 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2134 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2114, %2133)\n", " %2135 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2136 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.95, %2135)\n", " %2137 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2128, %4153, %2132, %2134, %2136) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.243 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.239, %2137) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.247 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.243) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2140 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2141 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2122, %2140)\n", " %2142 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2143 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2125, %2142)\n", " %2144 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2145 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2110, %2144)\n", " %2146 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2141, %2143, %2145) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.23 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.247, %2146) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2148 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2087) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2149 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2150 : Long(device=cpu) = onnx::Gather[axis=0](%2148, %2149) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2151 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2087) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2152 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2153 : Long(device=cpu) = onnx::Gather[axis=0](%2151, %2152) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2154 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2087) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2155 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2156 : Long(device=cpu) = onnx::Gather[axis=0](%2154, %2155) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2157 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2087) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2158 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2159 : Long(device=cpu) = onnx::Gather[axis=0](%2157, %2158) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2160 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2161 : Long(device=cpu) = onnx::Div(%2150, %2160) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2162 : Long(device=cpu) = onnx::Cast[to=7](%2161) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2163 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2162) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2164 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2165 : Long(device=cpu) = onnx::Div(%2153, %2164) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2166 : Long(device=cpu) = onnx::Cast[to=7](%2165) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2167 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2166) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2168 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2169 : Long(device=cpu) = onnx::Div(%2156, %2168) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2170 : Long(device=cpu) = onnx::Cast[to=7](%2169) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2171 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2170) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2172 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2173 : Long(device=cpu) = onnx::Div(%2159, %2172) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2174 : Long(device=cpu) = onnx::Cast[to=7](%2173) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.99 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2174) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2176 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2177 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2163, %2176) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2178 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2179 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2177, %2178) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2180 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2181 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2171, %2180) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2182 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2181, %inferred_length.99) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2184 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2185 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2163, %2184)\n", " %2188 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2189 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2167, %2188)\n", " %2190 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2191 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2171, %2190)\n", " %2192 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2193 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.99, %2192)\n", " %2194 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2185, %4154, %2189, %2191, %2193) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.251 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2087, %2194) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.255 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.251) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2197 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2198 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2179, %2197)\n", " %2199 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2200 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2182, %2199)\n", " %2201 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2202 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2167, %2201)\n", " %2203 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2198, %2200, %2202) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.23 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.255, %2203) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2205 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2090) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2206 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2207 : Long(device=cpu) = onnx::Gather[axis=0](%2205, %2206) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2208 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2090) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2209 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2210 : Long(device=cpu) = onnx::Gather[axis=0](%2208, %2209) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2211 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2090) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2212 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2213 : Long(device=cpu) = onnx::Gather[axis=0](%2211, %2212) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2214 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2090) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2215 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2216 : Long(device=cpu) = onnx::Gather[axis=0](%2214, %2215) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2217 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2218 : Long(device=cpu) = onnx::Div(%2207, %2217) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2219 : Long(device=cpu) = onnx::Cast[to=7](%2218) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2220 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2219) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2221 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2222 : Long(device=cpu) = onnx::Div(%2210, %2221) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2223 : Long(device=cpu) = onnx::Cast[to=7](%2222) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2224 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2223) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2225 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2226 : Long(device=cpu) = onnx::Div(%2213, %2225) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2227 : Long(device=cpu) = onnx::Cast[to=7](%2226) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2228 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2227) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2229 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2230 : Long(device=cpu) = onnx::Div(%2216, %2229) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2231 : Long(device=cpu) = onnx::Cast[to=7](%2230) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.103 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2231) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2233 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2234 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2220, %2233) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2235 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2236 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2234, %2235) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2237 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2238 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2228, %2237) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2239 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2238, %inferred_length.103) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2241 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2242 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2220, %2241)\n", " %2245 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2246 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2224, %2245)\n", " %2247 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2248 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2228, %2247)\n", " %2249 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2250 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.103, %2249)\n", " %2251 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2242, %4155, %2246, %2248, %2250) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.259 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2090, %2251) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.263 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.259) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2254 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2255 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2236, %2254)\n", " %2256 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2257 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2239, %2256)\n", " %2258 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2259 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2224, %2258)\n", " %2260 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2255, %2257, %2259) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.23 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.263, %2260) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2262 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.23, %k.23) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2263 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.316 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%2262, %2263) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.320 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.316) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.267 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.320, %v.23) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2267 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.267) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2268 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2269 : Long(device=cpu) = onnx::Gather[axis=0](%2267, %2268) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2270 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.267) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2271 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2272 : Long(device=cpu) = onnx::Gather[axis=0](%2270, %2271) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2273 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.267) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2274 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2275 : Long(device=cpu) = onnx::Gather[axis=0](%2273, %2274) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2276 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2277 : Long(device=cpu) = onnx::Div(%2269, %2276) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2278 : Long(device=cpu) = onnx::Cast[to=7](%2277) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2279 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2278) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2280 : Long(device=cpu) = onnx::Div(%2272, %2070) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2281 : Long(device=cpu) = onnx::Cast[to=7](%2280) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2282 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2281) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2283 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2284 : Long(device=cpu) = onnx::Div(%2275, %2283) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2285 : Long(device=cpu) = onnx::Cast[to=7](%2284) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.107 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2285) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2287 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %2288 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.107, %2287) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2290 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2291 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2279, %2290)\n", " %2294 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2295 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2282, %2294)\n", " %2296 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2297 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2070, %2296)\n", " %2298 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2299 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.107, %2298)\n", " %2300 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2291, %4156, %2295, %2297, %2299) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.271 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.267, %2300) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.275 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.271) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2303 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2304 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2279, %2303)\n", " %2305 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2306 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2288, %2305)\n", " %2307 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2308 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2282, %2307)\n", " %2309 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2310 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2070, %2309)\n", " %2311 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%2304, %2306, %2308, %2310) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.324 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.275, %2311) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.328 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.324, %layers.2.2.layers.3.0.fn.to_out.0.weight, %layers.2.2.layers.3.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2314 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.328, %2055) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %2315 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2314) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2316 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2314, %2315) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2317 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2316, %2316) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2318 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2317) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2319 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2314) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2320 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2314, %2319) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2321 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2322 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2318, %2321) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2323 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2322) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2324 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2320, %2323) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2325 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2324, %layers.2.2.layers.3.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.332 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2325, %layers.2.2.layers.3.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2327 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.332, %layers.2.2.layers.3.1.fn.net.0.weight, %layers.2.2.layers.3.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2328 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2329 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%2327, %2328) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2330 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%2329) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2331 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2332 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%2330, %2331) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2333 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%2327, %2332) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2334 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.336 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2333, %2334) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.340 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.336, %layers.2.2.layers.3.1.fn.net.3.weight, %layers.2.2.layers.3.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2337 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.340, %2314) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %2338 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2337) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2339 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2337, %2338) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2340 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2339, %2339) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2341 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2340) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2342 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2337) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2343 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2337, %2342) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2344 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2345 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2341, %2344) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2346 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2345) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2347 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2343, %2346) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2348 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2347, %layers.2.2.layers.4.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.344 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2348, %layers.2.2.layers.4.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2350 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.344) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2351 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2352 : Long(device=cpu) = onnx::Gather[axis=0](%2350, %2351) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.352 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.344, %4094, %4095) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.279 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.352, %layers.2.2.layers.4.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.360 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.344, %4097, %4098) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2358 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.360, %layers.2.2.layers.4.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2359 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2358) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2360 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2361 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%2359, %2360) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2362 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2363 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2364 : Long(1, strides=[1], device=cpu) = onnx::Add(%2361, %2363) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2365 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2366 : Long(1, strides=[1], device=cpu) = onnx::Div(%2364, %2365) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2367 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2368 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2366, %2367) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2369 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2358, %2362, %2368, %2360) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2370 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2371 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2366, %2370) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2372 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2358, %2368, %2371, %2360) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2373 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.279) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2374 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2375 : Long(device=cpu) = onnx::Gather[axis=0](%2373, %2374) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2376 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.279) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2377 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2378 : Long(device=cpu) = onnx::Gather[axis=0](%2376, %2377) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2379 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.279) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2380 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2381 : Long(device=cpu) = onnx::Gather[axis=0](%2379, %2380) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2382 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.279) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2383 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2384 : Long(device=cpu) = onnx::Gather[axis=0](%2382, %2383) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2385 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2386 : Long(device=cpu) = onnx::Div(%2375, %2385) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2387 : Long(device=cpu) = onnx::Cast[to=7](%2386) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2388 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2389 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2390 : Long(device=cpu) = onnx::Div(%2378, %2389) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2391 : Long(device=cpu) = onnx::Cast[to=7](%2390) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2392 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2391) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2393 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2394 : Long(device=cpu) = onnx::Div(%2381, %2393) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2395 : Long(device=cpu) = onnx::Cast[to=7](%2394) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2396 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2395) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2397 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2398 : Long(device=cpu) = onnx::Div(%2384, %2397) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2399 : Long(device=cpu) = onnx::Cast[to=7](%2398) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.111 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2401 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2402 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2388, %2401) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2403 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2404 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2402, %2403) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2405 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2406 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2396, %2405) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2407 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2406, %inferred_length.111) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2409 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2410 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2388, %2409)\n", " %2413 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2414 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2392, %2413)\n", " %2415 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2416 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2396, %2415)\n", " %2417 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2418 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.111, %2417)\n", " %2419 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2410, %4157, %2414, %2416, %2418) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.283 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.279, %2419) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.287 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.283) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2422 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2423 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2404, %2422)\n", " %2424 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2425 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2407, %2424)\n", " %2426 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2427 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2392, %2426)\n", " %2428 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2423, %2425, %2427) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.27 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.287, %2428) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2430 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2369) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2431 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2432 : Long(device=cpu) = onnx::Gather[axis=0](%2430, %2431) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2433 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2369) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2434 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2435 : Long(device=cpu) = onnx::Gather[axis=0](%2433, %2434) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2436 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2369) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2437 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2438 : Long(device=cpu) = onnx::Gather[axis=0](%2436, %2437) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2439 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2369) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2440 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2441 : Long(device=cpu) = onnx::Gather[axis=0](%2439, %2440) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2442 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2443 : Long(device=cpu) = onnx::Div(%2432, %2442) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2444 : Long(device=cpu) = onnx::Cast[to=7](%2443) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2445 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2444) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2446 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2447 : Long(device=cpu) = onnx::Div(%2435, %2446) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2448 : Long(device=cpu) = onnx::Cast[to=7](%2447) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2449 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2448) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2450 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2451 : Long(device=cpu) = onnx::Div(%2438, %2450) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2452 : Long(device=cpu) = onnx::Cast[to=7](%2451) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2453 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2452) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2454 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2455 : Long(device=cpu) = onnx::Div(%2441, %2454) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2456 : Long(device=cpu) = onnx::Cast[to=7](%2455) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.115 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2456) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2458 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2459 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2445, %2458) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2460 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2461 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2459, %2460) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2462 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2463 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2453, %2462) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2464 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2463, %inferred_length.115) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2466 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2467 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2445, %2466)\n", " %2470 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2471 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2449, %2470)\n", " %2472 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2473 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2453, %2472)\n", " %2474 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2475 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.115, %2474)\n", " %2476 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2467, %4158, %2471, %2473, %2475) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.291 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2369, %2476) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.295 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.291) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2479 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2480 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2461, %2479)\n", " %2481 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2482 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2464, %2481)\n", " %2483 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2484 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2449, %2483)\n", " %2485 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2480, %2482, %2484) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.27 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.295, %2485) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2487 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2372) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2488 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2489 : Long(device=cpu) = onnx::Gather[axis=0](%2487, %2488) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2490 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2372) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2491 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2492 : Long(device=cpu) = onnx::Gather[axis=0](%2490, %2491) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2493 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2372) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2494 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2495 : Long(device=cpu) = onnx::Gather[axis=0](%2493, %2494) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2496 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2372) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2497 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2498 : Long(device=cpu) = onnx::Gather[axis=0](%2496, %2497) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2499 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2500 : Long(device=cpu) = onnx::Div(%2489, %2499) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2501 : Long(device=cpu) = onnx::Cast[to=7](%2500) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2502 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2501) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2503 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2504 : Long(device=cpu) = onnx::Div(%2492, %2503) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2505 : Long(device=cpu) = onnx::Cast[to=7](%2504) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2506 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2505) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2507 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2508 : Long(device=cpu) = onnx::Div(%2495, %2507) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2509 : Long(device=cpu) = onnx::Cast[to=7](%2508) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2510 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2509) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2511 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2512 : Long(device=cpu) = onnx::Div(%2498, %2511) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2513 : Long(device=cpu) = onnx::Cast[to=7](%2512) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.119 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2513) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2515 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2516 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2502, %2515) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2517 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2518 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2516, %2517) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2519 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2520 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2510, %2519) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2521 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2520, %inferred_length.119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2523 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2524 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2502, %2523)\n", " %2527 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2528 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2506, %2527)\n", " %2529 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2530 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2510, %2529)\n", " %2531 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2532 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.119, %2531)\n", " %2533 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2524, %4159, %2528, %2530, %2532) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.299 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2372, %2533) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.303 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.299) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2536 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2537 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2518, %2536)\n", " %2538 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2539 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2521, %2538)\n", " %2540 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2541 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2506, %2540)\n", " %2542 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2537, %2539, %2541) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.27 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.303, %2542) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2544 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.27, %k.27) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2545 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.364 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%2544, %2545) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.368 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.364) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.307 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.368, %v.27) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2549 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.307) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2550 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2551 : Long(device=cpu) = onnx::Gather[axis=0](%2549, %2550) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2552 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.307) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2553 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2554 : Long(device=cpu) = onnx::Gather[axis=0](%2552, %2553) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2555 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.307) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2556 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2557 : Long(device=cpu) = onnx::Gather[axis=0](%2555, %2556) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2558 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2559 : Long(device=cpu) = onnx::Div(%2551, %2558) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2560 : Long(device=cpu) = onnx::Cast[to=7](%2559) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2561 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2560) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2562 : Long(device=cpu) = onnx::Div(%2554, %2352) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2563 : Long(device=cpu) = onnx::Cast[to=7](%2562) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2564 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2563) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2565 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2566 : Long(device=cpu) = onnx::Div(%2557, %2565) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2567 : Long(device=cpu) = onnx::Cast[to=7](%2566) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.123 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2567) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2569 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %2570 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.123, %2569) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2572 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2573 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2561, %2572)\n", " %2576 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2577 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2564, %2576)\n", " %2578 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2579 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2352, %2578)\n", " %2580 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2581 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.123, %2580)\n", " %2582 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2573, %4160, %2577, %2579, %2581) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.311 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.307, %2582) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.315 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.311) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2585 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2586 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2561, %2585)\n", " %2587 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2588 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2570, %2587)\n", " %2589 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2590 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2564, %2589)\n", " %2591 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2592 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2352, %2591)\n", " %2593 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%2586, %2588, %2590, %2592) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.372 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.315, %2593) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.376 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.372, %layers.2.2.layers.4.0.fn.to_out.0.weight, %layers.2.2.layers.4.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2596 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.376, %2337) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %2597 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2596) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2598 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2596, %2597) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2599 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2598, %2598) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2600 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2599) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2601 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2596) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2602 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2596, %2601) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2603 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2604 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2600, %2603) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2605 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2604) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2606 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2602, %2605) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2607 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2606, %layers.2.2.layers.4.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.380 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2607, %layers.2.2.layers.4.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2609 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.380, %layers.2.2.layers.4.1.fn.net.0.weight, %layers.2.2.layers.4.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2610 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2611 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%2609, %2610) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2612 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%2611) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2613 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2614 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%2612, %2613) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2615 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%2609, %2614) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2616 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.384 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2615, %2616) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.388 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.384, %layers.2.2.layers.4.1.fn.net.3.weight, %layers.2.2.layers.4.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2619 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.388, %2596) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %2620 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2619) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2621 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2619, %2620) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2622 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2621, %2621) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2623 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2622) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2624 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2619) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2625 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2619, %2624) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2626 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2627 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2623, %2626) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2628 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2627) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2629 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2625, %2628) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2630 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2629, %layers.2.2.layers.5.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.392 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2630, %layers.2.2.layers.5.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2632 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.392) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2633 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2634 : Long(device=cpu) = onnx::Gather[axis=0](%2632, %2633) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.400 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.392, %4100, %4101) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.319 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.400, %layers.2.2.layers.5.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.408 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.392, %4103, %4104) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2640 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.408, %layers.2.2.layers.5.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2641 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2640) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2642 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2643 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%2641, %2642) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2644 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2645 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2646 : Long(1, strides=[1], device=cpu) = onnx::Add(%2643, %2645) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2647 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2648 : Long(1, strides=[1], device=cpu) = onnx::Div(%2646, %2647) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2649 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2650 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2648, %2649) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2651 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2640, %2644, %2650, %2642) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2652 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2653 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2648, %2652) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2654 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2640, %2650, %2653, %2642) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2655 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.319) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2656 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2657 : Long(device=cpu) = onnx::Gather[axis=0](%2655, %2656) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2658 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.319) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2659 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2660 : Long(device=cpu) = onnx::Gather[axis=0](%2658, %2659) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2661 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.319) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2662 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2663 : Long(device=cpu) = onnx::Gather[axis=0](%2661, %2662) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2664 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.319) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2665 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2666 : Long(device=cpu) = onnx::Gather[axis=0](%2664, %2665) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2667 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2668 : Long(device=cpu) = onnx::Div(%2657, %2667) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2669 : Long(device=cpu) = onnx::Cast[to=7](%2668) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2670 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2669) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2671 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2672 : Long(device=cpu) = onnx::Div(%2660, %2671) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2673 : Long(device=cpu) = onnx::Cast[to=7](%2672) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2674 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2673) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2675 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2676 : Long(device=cpu) = onnx::Div(%2663, %2675) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2677 : Long(device=cpu) = onnx::Cast[to=7](%2676) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2678 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2677) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2679 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2680 : Long(device=cpu) = onnx::Div(%2666, %2679) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2681 : Long(device=cpu) = onnx::Cast[to=7](%2680) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.127 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2681) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2683 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2684 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2670, %2683) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2685 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2686 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2684, %2685) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2687 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2688 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2678, %2687) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2689 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2688, %inferred_length.127) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2691 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2692 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2670, %2691)\n", " %2695 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2696 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2674, %2695)\n", " %2697 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2698 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2678, %2697)\n", " %2699 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2700 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.127, %2699)\n", " %2701 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2692, %4161, %2696, %2698, %2700) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.323 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.319, %2701) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.327 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.323) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2704 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2705 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2686, %2704)\n", " %2706 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2707 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2689, %2706)\n", " %2708 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2709 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2674, %2708)\n", " %2710 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2705, %2707, %2709) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.31 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.327, %2710) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2712 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2651) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2713 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2714 : Long(device=cpu) = onnx::Gather[axis=0](%2712, %2713) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2715 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2651) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2716 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2717 : Long(device=cpu) = onnx::Gather[axis=0](%2715, %2716) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2718 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2651) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2719 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2720 : Long(device=cpu) = onnx::Gather[axis=0](%2718, %2719) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2721 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2651) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2722 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2723 : Long(device=cpu) = onnx::Gather[axis=0](%2721, %2722) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2724 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2725 : Long(device=cpu) = onnx::Div(%2714, %2724) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2726 : Long(device=cpu) = onnx::Cast[to=7](%2725) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2727 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2726) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2728 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2729 : Long(device=cpu) = onnx::Div(%2717, %2728) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2730 : Long(device=cpu) = onnx::Cast[to=7](%2729) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2731 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2730) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2732 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2733 : Long(device=cpu) = onnx::Div(%2720, %2732) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2734 : Long(device=cpu) = onnx::Cast[to=7](%2733) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2735 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2734) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2736 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2737 : Long(device=cpu) = onnx::Div(%2723, %2736) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2738 : Long(device=cpu) = onnx::Cast[to=7](%2737) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.131 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2738) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2740 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2741 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2727, %2740) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2742 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2743 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2741, %2742) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2744 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2745 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2735, %2744) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2746 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2745, %inferred_length.131) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2748 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2749 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2727, %2748)\n", " %2752 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2753 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2731, %2752)\n", " %2754 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2755 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2735, %2754)\n", " %2756 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2757 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.131, %2756)\n", " %2758 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2749, %4162, %2753, %2755, %2757) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.331 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2651, %2758) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.335 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.331) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2761 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2762 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2743, %2761)\n", " %2763 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2764 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2746, %2763)\n", " %2765 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2766 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2731, %2765)\n", " %2767 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2762, %2764, %2766) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.31 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.335, %2767) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2769 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2654) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2770 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2771 : Long(device=cpu) = onnx::Gather[axis=0](%2769, %2770) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2772 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2654) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2773 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2774 : Long(device=cpu) = onnx::Gather[axis=0](%2772, %2773) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2775 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2654) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2776 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2777 : Long(device=cpu) = onnx::Gather[axis=0](%2775, %2776) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2778 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2654) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2779 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2780 : Long(device=cpu) = onnx::Gather[axis=0](%2778, %2779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2781 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2782 : Long(device=cpu) = onnx::Div(%2771, %2781) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2783 : Long(device=cpu) = onnx::Cast[to=7](%2782) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2784 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2783) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2785 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2786 : Long(device=cpu) = onnx::Div(%2774, %2785) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2787 : Long(device=cpu) = onnx::Cast[to=7](%2786) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2788 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2787) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2789 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2790 : Long(device=cpu) = onnx::Div(%2777, %2789) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2791 : Long(device=cpu) = onnx::Cast[to=7](%2790) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2792 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2791) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2793 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2794 : Long(device=cpu) = onnx::Div(%2780, %2793) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2795 : Long(device=cpu) = onnx::Cast[to=7](%2794) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.135 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2795) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2797 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2798 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2784, %2797) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2799 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2800 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2798, %2799) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2801 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2802 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2792, %2801) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2803 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2802, %inferred_length.135) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2805 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2806 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2784, %2805)\n", " %2809 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2810 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2788, %2809)\n", " %2811 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2812 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2792, %2811)\n", " %2813 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2814 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.135, %2813)\n", " %2815 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2806, %4163, %2810, %2812, %2814) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.339 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2654, %2815) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.343 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.339) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2818 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2819 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2800, %2818)\n", " %2820 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2821 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2803, %2820)\n", " %2822 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2823 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2788, %2822)\n", " %2824 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2819, %2821, %2823) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.31 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.343, %2824) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2826 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.31, %k.31) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2827 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.412 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%2826, %2827) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.416 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.412) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.347 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.416, %v.31) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %2831 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.347) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2832 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2833 : Long(device=cpu) = onnx::Gather[axis=0](%2831, %2832) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2834 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.347) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2835 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2836 : Long(device=cpu) = onnx::Gather[axis=0](%2834, %2835) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2837 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.347) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2838 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2839 : Long(device=cpu) = onnx::Gather[axis=0](%2837, %2838) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2840 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2841 : Long(device=cpu) = onnx::Div(%2833, %2840) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2842 : Long(device=cpu) = onnx::Cast[to=7](%2841) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2843 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2842) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2844 : Long(device=cpu) = onnx::Div(%2836, %2634) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2845 : Long(device=cpu) = onnx::Cast[to=7](%2844) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2846 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2845) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2847 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2848 : Long(device=cpu) = onnx::Div(%2839, %2847) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2849 : Long(device=cpu) = onnx::Cast[to=7](%2848) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.139 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2849) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2851 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %2852 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.139, %2851) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2854 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2855 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2843, %2854)\n", " %2858 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2859 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2846, %2858)\n", " %2860 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2861 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2634, %2860)\n", " %2862 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2863 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.139, %2862)\n", " %2864 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2855, %4164, %2859, %2861, %2863) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.351 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.347, %2864) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.355 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.351) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2867 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2868 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2843, %2867)\n", " %2869 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2870 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2852, %2869)\n", " %2871 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2872 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2846, %2871)\n", " %2873 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2874 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2634, %2873)\n", " %2875 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%2868, %2870, %2872, %2874) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.420 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.355, %2875) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.424 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.420, %layers.2.2.layers.5.0.fn.to_out.0.weight, %layers.2.2.layers.5.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2878 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.424, %2619) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %2879 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2878) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2880 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2878, %2879) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2881 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2880, %2880) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2882 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2881) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2883 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2878) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2884 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2878, %2883) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2885 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2886 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2882, %2885) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2887 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2886) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2888 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2884, %2887) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2889 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2888, %layers.2.2.layers.5.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.428 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2889, %layers.2.2.layers.5.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2891 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.428, %layers.2.2.layers.5.1.fn.net.0.weight, %layers.2.2.layers.5.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2892 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2893 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%2891, %2892) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2894 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%2893) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2895 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2896 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%2894, %2895) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2897 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%2891, %2896) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %2898 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.432 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2897, %2898) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.436 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.432, %layers.2.2.layers.5.1.fn.net.3.weight, %layers.2.2.layers.5.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2901 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.436, %2878) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %2902 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2901) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2903 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%2901, %2902) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2904 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%2903, %2903) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2905 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2904) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %2906 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%2901) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %2907 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%2901, %2906) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2908 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %2909 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2905, %2908) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2910 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%2909) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2911 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%2907, %2910) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2912 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%2911, %layers.2.2.layers.6.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.440 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%2912, %layers.2.2.layers.6.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %2914 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.440) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2915 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %2916 : Long(device=cpu) = onnx::Gather[axis=0](%2914, %2915) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.448 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.440, %4106, %4107) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.359 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.448, %layers.2.2.layers.6.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.456 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.440, %4109, %4110) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2922 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.456, %layers.2.2.layers.6.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %2923 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2922) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2924 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2925 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%2923, %2924) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2926 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2927 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2928 : Long(1, strides=[1], device=cpu) = onnx::Add(%2925, %2927) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2929 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2930 : Long(1, strides=[1], device=cpu) = onnx::Div(%2928, %2929) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2931 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2932 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2930, %2931) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2933 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2922, %2926, %2932, %2924) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2934 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2935 : Long(1, strides=[1], device=cpu) = onnx::Mul(%2930, %2934) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2936 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%2922, %2932, %2935, %2924) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %2937 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.359) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2938 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2939 : Long(device=cpu) = onnx::Gather[axis=0](%2937, %2938) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2940 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.359) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2941 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2942 : Long(device=cpu) = onnx::Gather[axis=0](%2940, %2941) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2943 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.359) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2944 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2945 : Long(device=cpu) = onnx::Gather[axis=0](%2943, %2944) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2946 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.359) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2947 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2948 : Long(device=cpu) = onnx::Gather[axis=0](%2946, %2947) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2949 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2950 : Long(device=cpu) = onnx::Div(%2939, %2949) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2951 : Long(device=cpu) = onnx::Cast[to=7](%2950) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2952 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2951) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2953 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2954 : Long(device=cpu) = onnx::Div(%2942, %2953) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2955 : Long(device=cpu) = onnx::Cast[to=7](%2954) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2956 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2955) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2957 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2958 : Long(device=cpu) = onnx::Div(%2945, %2957) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2959 : Long(device=cpu) = onnx::Cast[to=7](%2958) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2960 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2959) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2961 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2962 : Long(device=cpu) = onnx::Div(%2948, %2961) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2963 : Long(device=cpu) = onnx::Cast[to=7](%2962) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.143 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%2963) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %2965 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2966 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2952, %2965) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2967 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2968 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2966, %2967) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2969 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %2970 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2960, %2969) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2971 : Long(requires_grad=0, device=cpu) = onnx::Mul(%2970, %inferred_length.143) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %2973 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2974 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2952, %2973)\n", " %2977 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2978 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2956, %2977)\n", " %2979 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2980 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2960, %2979)\n", " %2981 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2982 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.143, %2981)\n", " %2983 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%2974, %4165, %2978, %2980, %2982) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.363 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.359, %2983) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.367 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.363) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %2986 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2987 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2968, %2986)\n", " %2988 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2989 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2971, %2988)\n", " %2990 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %2991 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2956, %2990)\n", " %2992 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%2987, %2989, %2991) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.35 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.367, %2992) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %2994 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2933) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2995 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2996 : Long(device=cpu) = onnx::Gather[axis=0](%2994, %2995) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2997 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2933) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2998 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %2999 : Long(device=cpu) = onnx::Gather[axis=0](%2997, %2998) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3000 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2933) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3001 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3002 : Long(device=cpu) = onnx::Gather[axis=0](%3000, %3001) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3003 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2933) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3004 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3005 : Long(device=cpu) = onnx::Gather[axis=0](%3003, %3004) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3006 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3007 : Long(device=cpu) = onnx::Div(%2996, %3006) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3008 : Long(device=cpu) = onnx::Cast[to=7](%3007) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3009 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3008) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3010 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3011 : Long(device=cpu) = onnx::Div(%2999, %3010) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3012 : Long(device=cpu) = onnx::Cast[to=7](%3011) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3013 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3012) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3014 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3015 : Long(device=cpu) = onnx::Div(%3002, %3014) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3016 : Long(device=cpu) = onnx::Cast[to=7](%3015) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3017 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3016) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3018 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3019 : Long(device=cpu) = onnx::Div(%3005, %3018) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3020 : Long(device=cpu) = onnx::Cast[to=7](%3019) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.147 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3020) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3022 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3023 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3009, %3022) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3024 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3025 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3023, %3024) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3026 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3027 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3017, %3026) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3028 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3027, %inferred_length.147) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3030 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3031 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3009, %3030)\n", " %3034 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3035 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3013, %3034)\n", " %3036 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3037 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3017, %3036)\n", " %3038 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3039 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.147, %3038)\n", " %3040 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3031, %4166, %3035, %3037, %3039) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.371 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2933, %3040) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.375 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.371) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3043 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3044 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3025, %3043)\n", " %3045 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3046 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3028, %3045)\n", " %3047 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3048 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3013, %3047)\n", " %3049 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3044, %3046, %3048) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.35 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.375, %3049) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3051 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2936) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3052 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3053 : Long(device=cpu) = onnx::Gather[axis=0](%3051, %3052) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3054 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2936) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3055 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3056 : Long(device=cpu) = onnx::Gather[axis=0](%3054, %3055) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3057 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2936) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3058 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3059 : Long(device=cpu) = onnx::Gather[axis=0](%3057, %3058) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3060 : Long(4, strides=[1], device=cpu) = onnx::Shape(%2936) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3061 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3062 : Long(device=cpu) = onnx::Gather[axis=0](%3060, %3061) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3063 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3064 : Long(device=cpu) = onnx::Div(%3053, %3063) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3065 : Long(device=cpu) = onnx::Cast[to=7](%3064) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3066 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3065) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3067 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3068 : Long(device=cpu) = onnx::Div(%3056, %3067) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3069 : Long(device=cpu) = onnx::Cast[to=7](%3068) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3070 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3069) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3071 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3072 : Long(device=cpu) = onnx::Div(%3059, %3071) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3073 : Long(device=cpu) = onnx::Cast[to=7](%3072) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3074 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3073) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3075 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3076 : Long(device=cpu) = onnx::Div(%3062, %3075) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3077 : Long(device=cpu) = onnx::Cast[to=7](%3076) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.151 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3077) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3079 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3080 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3066, %3079) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3081 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3082 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3080, %3081) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3083 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3084 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3074, %3083) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3085 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3084, %inferred_length.151) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3087 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3088 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3066, %3087)\n", " %3091 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3092 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3070, %3091)\n", " %3093 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3094 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3074, %3093)\n", " %3095 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3096 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.151, %3095)\n", " %3097 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3088, %4167, %3092, %3094, %3096) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.379 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%2936, %3097) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.383 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.379) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3100 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3101 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3082, %3100)\n", " %3102 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3103 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3085, %3102)\n", " %3104 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3105 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3070, %3104)\n", " %3106 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3101, %3103, %3105) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.35 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.383, %3106) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3108 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.35, %k.35) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3109 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.460 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%3108, %3109) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.464 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.460) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.387 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.464, %v.35) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3113 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3114 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3115 : Long(device=cpu) = onnx::Gather[axis=0](%3113, %3114) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3116 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3117 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3118 : Long(device=cpu) = onnx::Gather[axis=0](%3116, %3117) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3119 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3120 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3121 : Long(device=cpu) = onnx::Gather[axis=0](%3119, %3120) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3122 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3123 : Long(device=cpu) = onnx::Div(%3115, %3122) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3124 : Long(device=cpu) = onnx::Cast[to=7](%3123) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3125 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3124) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3126 : Long(device=cpu) = onnx::Div(%3118, %2916) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3127 : Long(device=cpu) = onnx::Cast[to=7](%3126) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3128 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3127) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3129 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3130 : Long(device=cpu) = onnx::Div(%3121, %3129) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3131 : Long(device=cpu) = onnx::Cast[to=7](%3130) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.155 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3131) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3133 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %3134 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.155, %3133) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3136 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3137 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3125, %3136)\n", " %3140 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3141 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3128, %3140)\n", " %3142 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3143 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2916, %3142)\n", " %3144 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3145 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.155, %3144)\n", " %3146 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3137, %4168, %3141, %3143, %3145) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.391 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.387, %3146) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.395 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.391) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3149 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3150 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3125, %3149)\n", " %3151 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3152 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3134, %3151)\n", " %3153 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3154 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3128, %3153)\n", " %3155 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3156 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%2916, %3155)\n", " %3157 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%3150, %3152, %3154, %3156) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.468 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.395, %3157) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.472 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.468, %layers.2.2.layers.6.0.fn.to_out.0.weight, %layers.2.2.layers.6.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3160 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.472, %2901) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %3161 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3160) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3162 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3160, %3161) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3163 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3162, %3162) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3164 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3163) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3165 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3160) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3166 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3160, %3165) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3167 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3168 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3164, %3167) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3169 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3168) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3170 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3166, %3169) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3171 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3170, %layers.2.2.layers.6.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.476 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3171, %layers.2.2.layers.6.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3173 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.476, %layers.2.2.layers.6.1.fn.net.0.weight, %layers.2.2.layers.6.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3174 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3175 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%3173, %3174) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3176 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%3175) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3177 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3178 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%3176, %3177) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3179 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%3173, %3178) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3180 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.480 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3179, %3180) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.484 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.480, %layers.2.2.layers.6.1.fn.net.3.weight, %layers.2.2.layers.6.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3183 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.484, %3160) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %3184 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3183) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3185 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3183, %3184) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3186 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3185, %3185) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3187 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3186) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3188 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3183) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3189 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3183, %3188) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3190 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3191 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3187, %3190) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3192 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3191) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3193 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3189, %3192) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3194 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3193, %layers.2.2.layers.7.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.488 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3194, %layers.2.2.layers.7.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3196 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.488) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3197 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3198 : Long(device=cpu) = onnx::Gather[axis=0](%3196, %3197) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.496 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.488, %4112, %4113) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.399 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.496, %layers.2.2.layers.7.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.504 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.488, %4115, %4116) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3204 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.504, %layers.2.2.layers.7.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3205 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3204) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3206 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3207 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%3205, %3206) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3208 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3209 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3210 : Long(1, strides=[1], device=cpu) = onnx::Add(%3207, %3209) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3211 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3212 : Long(1, strides=[1], device=cpu) = onnx::Div(%3210, %3211) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3213 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3214 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3212, %3213) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3215 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3204, %3208, %3214, %3206) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3216 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3217 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3212, %3216) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3218 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3204, %3214, %3217, %3206) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3219 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3220 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3221 : Long(device=cpu) = onnx::Gather[axis=0](%3219, %3220) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3222 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3223 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3224 : Long(device=cpu) = onnx::Gather[axis=0](%3222, %3223) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3225 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3226 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3227 : Long(device=cpu) = onnx::Gather[axis=0](%3225, %3226) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3228 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3229 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3230 : Long(device=cpu) = onnx::Gather[axis=0](%3228, %3229) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3231 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3232 : Long(device=cpu) = onnx::Div(%3221, %3231) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3233 : Long(device=cpu) = onnx::Cast[to=7](%3232) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3234 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3233) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3235 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3236 : Long(device=cpu) = onnx::Div(%3224, %3235) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3237 : Long(device=cpu) = onnx::Cast[to=7](%3236) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3238 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3237) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3239 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3240 : Long(device=cpu) = onnx::Div(%3227, %3239) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3241 : Long(device=cpu) = onnx::Cast[to=7](%3240) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3242 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3241) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3243 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3244 : Long(device=cpu) = onnx::Div(%3230, %3243) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3245 : Long(device=cpu) = onnx::Cast[to=7](%3244) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.159 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3245) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3247 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3248 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3234, %3247) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3249 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3250 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3248, %3249) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3251 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3252 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3242, %3251) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3253 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3252, %inferred_length.159) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3255 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3256 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3234, %3255)\n", " %3259 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3260 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3238, %3259)\n", " %3261 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3262 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3242, %3261)\n", " %3263 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3264 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.159, %3263)\n", " %3265 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3256, %4169, %3260, %3262, %3264) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.403 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.399, %3265) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.407 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.403) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3268 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3269 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3250, %3268)\n", " %3270 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3271 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3253, %3270)\n", " %3272 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3273 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3238, %3272)\n", " %3274 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3269, %3271, %3273) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.39 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.407, %3274) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3276 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3215) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3277 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3278 : Long(device=cpu) = onnx::Gather[axis=0](%3276, %3277) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3279 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3215) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3280 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3281 : Long(device=cpu) = onnx::Gather[axis=0](%3279, %3280) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3282 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3215) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3283 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3284 : Long(device=cpu) = onnx::Gather[axis=0](%3282, %3283) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3285 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3215) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3286 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3287 : Long(device=cpu) = onnx::Gather[axis=0](%3285, %3286) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3288 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3289 : Long(device=cpu) = onnx::Div(%3278, %3288) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3290 : Long(device=cpu) = onnx::Cast[to=7](%3289) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3291 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3290) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3292 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3293 : Long(device=cpu) = onnx::Div(%3281, %3292) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3294 : Long(device=cpu) = onnx::Cast[to=7](%3293) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3295 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3294) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3296 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3297 : Long(device=cpu) = onnx::Div(%3284, %3296) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3298 : Long(device=cpu) = onnx::Cast[to=7](%3297) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3299 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3298) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3300 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3301 : Long(device=cpu) = onnx::Div(%3287, %3300) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3302 : Long(device=cpu) = onnx::Cast[to=7](%3301) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.163 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3302) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3304 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3305 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3291, %3304) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3306 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3307 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3305, %3306) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3308 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3309 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3299, %3308) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3310 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3309, %inferred_length.163) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3312 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3313 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3291, %3312)\n", " %3316 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3317 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3295, %3316)\n", " %3318 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3319 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3299, %3318)\n", " %3320 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3321 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.163, %3320)\n", " %3322 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3313, %4170, %3317, %3319, %3321) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.411 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3215, %3322) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.415 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.411) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3325 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3326 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3307, %3325)\n", " %3327 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3328 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3310, %3327)\n", " %3329 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3330 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3295, %3329)\n", " %3331 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3326, %3328, %3330) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.39 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.415, %3331) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3333 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3218) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3334 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3335 : Long(device=cpu) = onnx::Gather[axis=0](%3333, %3334) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3336 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3218) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3337 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3338 : Long(device=cpu) = onnx::Gather[axis=0](%3336, %3337) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3339 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3218) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3340 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3341 : Long(device=cpu) = onnx::Gather[axis=0](%3339, %3340) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3342 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3218) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3343 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3344 : Long(device=cpu) = onnx::Gather[axis=0](%3342, %3343) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3345 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3346 : Long(device=cpu) = onnx::Div(%3335, %3345) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3347 : Long(device=cpu) = onnx::Cast[to=7](%3346) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3348 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3347) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3349 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3350 : Long(device=cpu) = onnx::Div(%3338, %3349) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3351 : Long(device=cpu) = onnx::Cast[to=7](%3350) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3352 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3351) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3353 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3354 : Long(device=cpu) = onnx::Div(%3341, %3353) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3355 : Long(device=cpu) = onnx::Cast[to=7](%3354) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3356 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3355) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3357 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3358 : Long(device=cpu) = onnx::Div(%3344, %3357) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3359 : Long(device=cpu) = onnx::Cast[to=7](%3358) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.167 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3359) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3361 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3362 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3348, %3361) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3363 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3364 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3362, %3363) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3365 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3366 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3356, %3365) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3367 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3366, %inferred_length.167) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3369 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3370 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3348, %3369)\n", " %3373 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3374 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3352, %3373)\n", " %3375 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3376 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3356, %3375)\n", " %3377 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3378 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.167, %3377)\n", " %3379 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3370, %4171, %3374, %3376, %3378) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.419 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3218, %3379) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.423 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.419) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3382 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3383 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3364, %3382)\n", " %3384 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3385 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3367, %3384)\n", " %3386 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3387 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3352, %3386)\n", " %3388 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3383, %3385, %3387) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.39 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.423, %3388) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3390 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.39, %k.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3391 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.508 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%3390, %3391) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.512 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.508) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.427 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.512, %v.39) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3395 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.427) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3396 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3397 : Long(device=cpu) = onnx::Gather[axis=0](%3395, %3396) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3398 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.427) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3399 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3400 : Long(device=cpu) = onnx::Gather[axis=0](%3398, %3399) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3401 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.427) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3402 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3403 : Long(device=cpu) = onnx::Gather[axis=0](%3401, %3402) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3404 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3405 : Long(device=cpu) = onnx::Div(%3397, %3404) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3406 : Long(device=cpu) = onnx::Cast[to=7](%3405) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3407 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3406) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3408 : Long(device=cpu) = onnx::Div(%3400, %3198) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3409 : Long(device=cpu) = onnx::Cast[to=7](%3408) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3410 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3409) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3411 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3412 : Long(device=cpu) = onnx::Div(%3403, %3411) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3413 : Long(device=cpu) = onnx::Cast[to=7](%3412) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.171 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3413) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3415 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %3416 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.171, %3415) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3418 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3419 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3407, %3418)\n", " %3422 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3423 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3410, %3422)\n", " %3424 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3425 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3198, %3424)\n", " %3426 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3427 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.171, %3426)\n", " %3428 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3419, %4172, %3423, %3425, %3427) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.431 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.427, %3428) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.435 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.431) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3431 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3432 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3407, %3431)\n", " %3433 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3434 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3416, %3433)\n", " %3435 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3436 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3410, %3435)\n", " %3437 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3438 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3198, %3437)\n", " %3439 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%3432, %3434, %3436, %3438) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.516 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.435, %3439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.520 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.516, %layers.2.2.layers.7.0.fn.to_out.0.weight, %layers.2.2.layers.7.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3442 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.520, %3183) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %3443 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3442) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3444 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3442, %3443) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3445 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3444, %3444) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3446 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3445) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3447 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3442) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3448 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3442, %3447) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3449 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3450 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3446, %3449) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3451 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3450) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3452 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3448, %3451) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3453 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3452, %layers.2.2.layers.7.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.524 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3453, %layers.2.2.layers.7.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3455 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.524, %layers.2.2.layers.7.1.fn.net.0.weight, %layers.2.2.layers.7.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3456 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3457 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%3455, %3456) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3458 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%3457) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3459 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3460 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%3458, %3459) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3461 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%3455, %3460) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3462 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.528 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3461, %3462) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.532 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.528, %layers.2.2.layers.7.1.fn.net.3.weight, %layers.2.2.layers.7.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3465 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.532, %3442) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %3466 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3465) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3467 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3465, %3466) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3468 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3467, %3467) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3469 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3468) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3470 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3465) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3471 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3465, %3470) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3472 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3473 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3469, %3472) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3474 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3473) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3475 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3471, %3474) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3476 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3475, %layers.2.2.layers.8.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.536 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3476, %layers.2.2.layers.8.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3478 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.536) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3479 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3480 : Long(device=cpu) = onnx::Gather[axis=0](%3478, %3479) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.544 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.536, %4118, %4119) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.439 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.544, %layers.2.2.layers.8.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.552 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.536, %4121, %4122) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3486 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.552, %layers.2.2.layers.8.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3487 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3486) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3488 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3489 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%3487, %3488) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3490 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3491 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3492 : Long(1, strides=[1], device=cpu) = onnx::Add(%3489, %3491) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3493 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3494 : Long(1, strides=[1], device=cpu) = onnx::Div(%3492, %3493) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3495 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3496 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3494, %3495) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3497 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3486, %3490, %3496, %3488) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3498 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3499 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3494, %3498) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3500 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3486, %3496, %3499, %3488) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3501 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3502 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3503 : Long(device=cpu) = onnx::Gather[axis=0](%3501, %3502) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3504 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3505 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3506 : Long(device=cpu) = onnx::Gather[axis=0](%3504, %3505) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3507 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3508 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3509 : Long(device=cpu) = onnx::Gather[axis=0](%3507, %3508) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3510 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.439) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3511 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3512 : Long(device=cpu) = onnx::Gather[axis=0](%3510, %3511) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3513 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3514 : Long(device=cpu) = onnx::Div(%3503, %3513) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3515 : Long(device=cpu) = onnx::Cast[to=7](%3514) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3516 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3515) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3517 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3518 : Long(device=cpu) = onnx::Div(%3506, %3517) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3519 : Long(device=cpu) = onnx::Cast[to=7](%3518) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3520 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3519) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3521 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3522 : Long(device=cpu) = onnx::Div(%3509, %3521) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3523 : Long(device=cpu) = onnx::Cast[to=7](%3522) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3524 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3523) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3525 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3526 : Long(device=cpu) = onnx::Div(%3512, %3525) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3527 : Long(device=cpu) = onnx::Cast[to=7](%3526) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.175 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3527) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3529 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3530 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3516, %3529) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3531 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3532 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3530, %3531) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3533 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3534 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3524, %3533) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3535 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3534, %inferred_length.175) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3537 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3538 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3516, %3537)\n", " %3541 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3542 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3520, %3541)\n", " %3543 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3544 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3524, %3543)\n", " %3545 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3546 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.175, %3545)\n", " %3547 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3538, %4173, %3542, %3544, %3546) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.443 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.439, %3547) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.447 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.443) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3550 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3551 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3532, %3550)\n", " %3552 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3553 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3535, %3552)\n", " %3554 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3555 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3520, %3554)\n", " %3556 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3551, %3553, %3555) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.43 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.447, %3556) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3558 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3497) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3559 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3560 : Long(device=cpu) = onnx::Gather[axis=0](%3558, %3559) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3561 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3497) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3562 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3563 : Long(device=cpu) = onnx::Gather[axis=0](%3561, %3562) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3564 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3497) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3565 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3566 : Long(device=cpu) = onnx::Gather[axis=0](%3564, %3565) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3567 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3497) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3568 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3569 : Long(device=cpu) = onnx::Gather[axis=0](%3567, %3568) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3570 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3571 : Long(device=cpu) = onnx::Div(%3560, %3570) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3572 : Long(device=cpu) = onnx::Cast[to=7](%3571) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3573 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3572) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3574 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3575 : Long(device=cpu) = onnx::Div(%3563, %3574) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3576 : Long(device=cpu) = onnx::Cast[to=7](%3575) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3577 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3576) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3578 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3579 : Long(device=cpu) = onnx::Div(%3566, %3578) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3580 : Long(device=cpu) = onnx::Cast[to=7](%3579) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3581 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3580) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3582 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3583 : Long(device=cpu) = onnx::Div(%3569, %3582) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3584 : Long(device=cpu) = onnx::Cast[to=7](%3583) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.179 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3584) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3586 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3587 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3573, %3586) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3588 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3589 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3587, %3588) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3590 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3591 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3581, %3590) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3592 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3591, %inferred_length.179) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3594 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3595 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3573, %3594)\n", " %3598 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3599 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3577, %3598)\n", " %3600 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3601 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3581, %3600)\n", " %3602 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3603 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.179, %3602)\n", " %3604 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3595, %4174, %3599, %3601, %3603) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.451 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3497, %3604) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.455 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.451) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3607 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3608 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3589, %3607)\n", " %3609 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3610 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3592, %3609)\n", " %3611 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3612 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3577, %3611)\n", " %3613 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3608, %3610, %3612) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.43 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.455, %3613) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3615 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3500) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3616 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3617 : Long(device=cpu) = onnx::Gather[axis=0](%3615, %3616) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3618 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3500) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3619 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3620 : Long(device=cpu) = onnx::Gather[axis=0](%3618, %3619) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3621 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3500) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3622 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3623 : Long(device=cpu) = onnx::Gather[axis=0](%3621, %3622) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3624 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3500) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3625 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3626 : Long(device=cpu) = onnx::Gather[axis=0](%3624, %3625) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3627 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3628 : Long(device=cpu) = onnx::Div(%3617, %3627) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3629 : Long(device=cpu) = onnx::Cast[to=7](%3628) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3630 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3629) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3631 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3632 : Long(device=cpu) = onnx::Div(%3620, %3631) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3633 : Long(device=cpu) = onnx::Cast[to=7](%3632) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3634 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3633) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3635 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3636 : Long(device=cpu) = onnx::Div(%3623, %3635) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3637 : Long(device=cpu) = onnx::Cast[to=7](%3636) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3638 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3637) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3639 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3640 : Long(device=cpu) = onnx::Div(%3626, %3639) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3641 : Long(device=cpu) = onnx::Cast[to=7](%3640) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.183 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3641) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3643 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3644 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3630, %3643) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3645 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3646 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3644, %3645) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3647 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3648 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3638, %3647) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3649 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3648, %inferred_length.183) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3651 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3652 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3630, %3651)\n", " %3655 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3656 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3634, %3655)\n", " %3657 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3658 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3638, %3657)\n", " %3659 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3660 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.183, %3659)\n", " %3661 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3652, %4175, %3656, %3658, %3660) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.459 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3500, %3661) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.463 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.459) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3664 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3665 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3646, %3664)\n", " %3666 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3667 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3649, %3666)\n", " %3668 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3669 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3634, %3668)\n", " %3670 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3665, %3667, %3669) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.43 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.463, %3670) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3672 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.43, %k.43) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3673 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.556 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%3672, %3673) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.560 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.556) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.467 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.560, %v.43) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3677 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.467) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3678 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3679 : Long(device=cpu) = onnx::Gather[axis=0](%3677, %3678) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3680 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.467) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3681 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3682 : Long(device=cpu) = onnx::Gather[axis=0](%3680, %3681) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3683 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.467) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3684 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3685 : Long(device=cpu) = onnx::Gather[axis=0](%3683, %3684) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3686 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3687 : Long(device=cpu) = onnx::Div(%3679, %3686) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3688 : Long(device=cpu) = onnx::Cast[to=7](%3687) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3689 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3688) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3690 : Long(device=cpu) = onnx::Div(%3682, %3480) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3691 : Long(device=cpu) = onnx::Cast[to=7](%3690) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3692 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3691) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3693 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3694 : Long(device=cpu) = onnx::Div(%3685, %3693) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3695 : Long(device=cpu) = onnx::Cast[to=7](%3694) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.187 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3695) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3697 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %3698 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.187, %3697) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3700 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3701 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3689, %3700)\n", " %3704 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3705 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3692, %3704)\n", " %3706 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3707 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3480, %3706)\n", " %3708 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3709 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.187, %3708)\n", " %3710 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3701, %4176, %3705, %3707, %3709) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.471 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.467, %3710) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.475 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.471) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3713 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3714 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3689, %3713)\n", " %3715 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3716 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3698, %3715)\n", " %3717 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3718 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3692, %3717)\n", " %3719 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3720 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3480, %3719)\n", " %3721 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%3714, %3716, %3718, %3720) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.564 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.475, %3721) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.568 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.564, %layers.2.2.layers.8.0.fn.to_out.0.weight, %layers.2.2.layers.8.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3724 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.568, %3465) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %3725 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3724) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3726 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3724, %3725) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3727 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3726, %3726) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3728 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3727) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3729 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3724) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3730 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3724, %3729) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3731 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3732 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3728, %3731) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3733 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3732) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3734 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3730, %3733) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3735 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3734, %layers.2.2.layers.8.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.572 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3735, %layers.2.2.layers.8.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3737 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.572, %layers.2.2.layers.8.1.fn.net.0.weight, %layers.2.2.layers.8.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3738 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3739 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%3737, %3738) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3740 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%3739) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3741 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3742 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%3740, %3741) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3743 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%3737, %3742) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %3744 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.576 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3743, %3744) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.580 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.576, %layers.2.2.layers.8.1.fn.net.3.weight, %layers.2.2.layers.8.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3747 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.580, %3724) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %3748 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3747) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3749 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%3747, %3748) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3750 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%3749, %3749) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3751 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3750) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %3752 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%3747) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %3753 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%3747, %3752) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3754 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %3755 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3751, %3754) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3756 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%3755) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3757 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%3753, %3756) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3758 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%3757, %layers.2.2.layers.9.0.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.584 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%3758, %layers.2.2.layers.9.0.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %3760 : Long(4, strides=[1], device=cpu) = onnx::Shape(%input.584) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3761 : Long(device=cpu) = onnx::Constant[value={3}]() # /tmp/ipykernel_26455/952968418.py:89:0\n", " %3762 : Long(device=cpu) = onnx::Gather[axis=0](%3760, %3761) # /tmp/ipykernel_26455/952968418.py:89:0\n", " %input.592 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.584, %4124, %4125) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %tensor.479 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.592, %layers.2.2.layers.9.0.fn.to_q.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.600 : Float(*, 384, *, *, strides=[24576, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=384, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%input.584, %4127, %4128) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3768 : Float(*, 768, *, *, strides=[49152, 64, 8, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.600, %layers.2.2.layers.9.0.fn.to_kv.net.2.weight) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %3769 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3768) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3770 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3771 : Long(1, strides=[1], device=cpu) = onnx::Gather[axis=0](%3769, %3770) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3772 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3773 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3774 : Long(1, strides=[1], device=cpu) = onnx::Add(%3771, %3773) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3775 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3776 : Long(1, strides=[1], device=cpu) = onnx::Div(%3774, %3775) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3777 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3778 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3776, %3777) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3779 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3768, %3772, %3778, %3770) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3780 : Long(1, strides=[1], requires_grad=0, device=cpu) = onnx::Constant[value={2}]() # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3781 : Long(1, strides=[1], device=cpu) = onnx::Mul(%3776, %3780) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3782 : Float(*, *, *, *, strides=[49152, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Slice(%3768, %3778, %3781, %3770) # /tmp/ipykernel_26455/952968418.py:91:0\n", " %3783 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.479) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3784 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3785 : Long(device=cpu) = onnx::Gather[axis=0](%3783, %3784) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3786 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.479) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3787 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3788 : Long(device=cpu) = onnx::Gather[axis=0](%3786, %3787) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3789 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.479) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3790 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3791 : Long(device=cpu) = onnx::Gather[axis=0](%3789, %3790) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3792 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.479) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3793 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3794 : Long(device=cpu) = onnx::Gather[axis=0](%3792, %3793) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3795 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3796 : Long(device=cpu) = onnx::Div(%3785, %3795) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3797 : Long(device=cpu) = onnx::Cast[to=7](%3796) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3798 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3797) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3799 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3800 : Long(device=cpu) = onnx::Div(%3788, %3799) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3801 : Long(device=cpu) = onnx::Cast[to=7](%3800) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3802 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3801) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3803 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3804 : Long(device=cpu) = onnx::Div(%3791, %3803) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3805 : Long(device=cpu) = onnx::Cast[to=7](%3804) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3806 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3805) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3807 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3808 : Long(device=cpu) = onnx::Div(%3794, %3807) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3809 : Long(device=cpu) = onnx::Cast[to=7](%3808) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.191 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3809) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3811 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3812 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3798, %3811) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3813 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3814 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3812, %3813) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3815 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3816 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3806, %3815) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3817 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3816, %inferred_length.191) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3819 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3820 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3798, %3819)\n", " %3823 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3824 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3802, %3823)\n", " %3825 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3826 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3806, %3825)\n", " %3827 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3828 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.191, %3827)\n", " %3829 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3820, %4177, %3824, %3826, %3828) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.483 : Float(*, *, *, *, *, strides=[98304, 16384, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.479, %3829) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.487 : Float(*, *, *, *, *, strides=[98304, 16384, 16, 1, 256], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.483) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3832 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3833 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3814, %3832)\n", " %3834 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3835 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3817, %3834)\n", " %3836 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3837 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3802, %3836)\n", " %3838 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3833, %3835, %3837) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %q.47 : Float(*, *, *, strides=[16384, 1, 256], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.487, %3838) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3840 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3841 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3842 : Long(device=cpu) = onnx::Gather[axis=0](%3840, %3841) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3843 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3844 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3845 : Long(device=cpu) = onnx::Gather[axis=0](%3843, %3844) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3846 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3847 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3848 : Long(device=cpu) = onnx::Gather[axis=0](%3846, %3847) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3849 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3779) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3850 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3851 : Long(device=cpu) = onnx::Gather[axis=0](%3849, %3850) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3852 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3853 : Long(device=cpu) = onnx::Div(%3842, %3852) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3854 : Long(device=cpu) = onnx::Cast[to=7](%3853) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3855 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3854) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3856 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3857 : Long(device=cpu) = onnx::Div(%3845, %3856) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3858 : Long(device=cpu) = onnx::Cast[to=7](%3857) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3859 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3858) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3860 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3861 : Long(device=cpu) = onnx::Div(%3848, %3860) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3862 : Long(device=cpu) = onnx::Cast[to=7](%3861) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3863 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3862) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3864 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3865 : Long(device=cpu) = onnx::Div(%3851, %3864) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3866 : Long(device=cpu) = onnx::Cast[to=7](%3865) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.195 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3866) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3868 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3869 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3855, %3868) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3870 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3871 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3869, %3870) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3872 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3873 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3863, %3872) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3874 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3873, %inferred_length.195) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3876 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3877 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3855, %3876)\n", " %3880 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3881 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3859, %3880)\n", " %3882 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3883 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3863, %3882)\n", " %3884 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3885 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.195, %3884)\n", " %3886 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3877, %4178, %3881, %3883, %3885) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.491 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3779, %3886) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.495 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.491) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3889 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3890 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3871, %3889)\n", " %3891 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3892 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3874, %3891)\n", " %3893 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3894 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3859, %3893)\n", " %3895 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3890, %3892, %3894) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %k.47 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.495, %3895) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3897 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3782) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3898 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3899 : Long(device=cpu) = onnx::Gather[axis=0](%3897, %3898) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3900 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3782) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3901 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3902 : Long(device=cpu) = onnx::Gather[axis=0](%3900, %3901) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3903 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3782) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3904 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3905 : Long(device=cpu) = onnx::Gather[axis=0](%3903, %3904) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3906 : Long(4, strides=[1], device=cpu) = onnx::Shape(%3782) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3907 : Long(device=cpu) = onnx::Constant[value={3}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3908 : Long(device=cpu) = onnx::Gather[axis=0](%3906, %3907) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3909 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3910 : Long(device=cpu) = onnx::Div(%3899, %3909) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3911 : Long(device=cpu) = onnx::Cast[to=7](%3910) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3912 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3911) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3913 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3914 : Long(device=cpu) = onnx::Div(%3902, %3913) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3915 : Long(device=cpu) = onnx::Cast[to=7](%3914) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3916 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3915) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3917 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3918 : Long(device=cpu) = onnx::Div(%3905, %3917) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3919 : Long(device=cpu) = onnx::Cast[to=7](%3918) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3920 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3919) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3921 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3922 : Long(device=cpu) = onnx::Div(%3908, %3921) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3923 : Long(device=cpu) = onnx::Cast[to=7](%3922) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.199 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3923) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3925 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3926 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3912, %3925) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3927 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3928 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3926, %3927) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3929 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %3930 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3920, %3929) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3931 : Long(requires_grad=0, device=cpu) = onnx::Mul(%3930, %inferred_length.199) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3933 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3934 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3912, %3933)\n", " %3937 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3938 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3916, %3937)\n", " %3939 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3940 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3920, %3939)\n", " %3941 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3942 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.199, %3941)\n", " %3943 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3934, %4179, %3938, %3940, %3942) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.499 : Float(*, *, *, *, *, strides=[24576, 4096, 64, 8, 1], requires_grad=1, device=cpu) = onnx::Reshape(%3782, %3943) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.503 : Float(*, *, *, *, *, strides=[24576, 4096, 8, 1, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%tensor.499) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3946 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3947 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3928, %3946)\n", " %3948 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3949 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3931, %3948)\n", " %3950 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3951 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3916, %3950)\n", " %3952 : Long(3, strides=[1], device=cpu) = onnx::Concat[axis=0](%3947, %3949, %3951) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %v.47 : Float(*, *, *, strides=[4096, 1, 64], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.503, %3952) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %3954 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i d, b j d -> b i j\"](%q.47, %k.47) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3955 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.125}]()\n", " %input.604 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Mul(%3954, %3955) # /tmp/ipykernel_26455/952968418.py:94:0\n", " %input.608 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Softmax[axis=-1](%input.604) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1818:0\n", " %tensor.507 : Float(*, *, *, strides=[16384, 64, 1], requires_grad=1, device=cpu) = onnx::Einsum[equation=\"b i j, b j d -> b i d\"](%input.608, %v.47) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/functional.py:330:0\n", " %3959 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.507) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3960 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3961 : Long(device=cpu) = onnx::Gather[axis=0](%3959, %3960) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3962 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.507) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3963 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3964 : Long(device=cpu) = onnx::Gather[axis=0](%3962, %3963) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3965 : Long(3, strides=[1], device=cpu) = onnx::Shape(%tensor.507) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3966 : Long(device=cpu) = onnx::Constant[value={2}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3967 : Long(device=cpu) = onnx::Gather[axis=0](%3965, %3966) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:81:0\n", " %3968 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3969 : Long(device=cpu) = onnx::Div(%3961, %3968) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3970 : Long(device=cpu) = onnx::Cast[to=7](%3969) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3971 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3970) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3972 : Long(device=cpu) = onnx::Div(%3964, %3762) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3973 : Long(device=cpu) = onnx::Cast[to=7](%3972) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3974 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3973) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3975 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3976 : Long(device=cpu) = onnx::Div(%3967, %3975) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3977 : Long(device=cpu) = onnx::Cast[to=7](%3976) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %inferred_length.203 : Long(requires_grad=0, device=cpu) = onnx::Cast[to=7](%3977) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/_tensor.py:647:0\n", " %3979 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={6}]()\n", " %3980 : Long(requires_grad=0, device=cpu) = onnx::Mul(%inferred_length.203, %3979) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %3982 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3983 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3971, %3982)\n", " %3986 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3987 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3974, %3986)\n", " %3988 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3989 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3762, %3988)\n", " %3990 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3991 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%inferred_length.203, %3990)\n", " %3992 : Long(5, strides=[1], device=cpu) = onnx::Concat[axis=0](%3983, %4180, %3987, %3989, %3991) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.511 : Float(*, *, *, *, *, strides=[98304, 16384, 1024, 64, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.507, %3992) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %tensor.515 : Float(*, *, *, *, *, strides=[98304, 16384, 1, 1024, 64], requires_grad=1, device=cpu) = onnx::Transpose[perm=[0, 1, 4, 2, 3]](%tensor.511) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:355:0\n", " %3995 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3996 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3971, %3995)\n", " %3997 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %3998 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3980, %3997)\n", " %3999 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %4000 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3974, %3999)\n", " %4001 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %4002 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%3762, %4001)\n", " %4003 : Long(4, strides=[1], device=cpu) = onnx::Concat[axis=0](%3996, %3998, %4000, %4002) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.612 : Float(*, *, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.515, %4003) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_backends.py:84:0\n", " %input.616 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.612, %layers.2.2.layers.9.0.fn.to_out.0.weight, %layers.2.2.layers.9.0.fn.to_out.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %4006 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.616, %3747) # /tmp/ipykernel_26455/952968418.py:114:0\n", " %4007 : Float(*, 1, *, *, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%4006) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %4008 : Float(*, 384, *, *, device=cpu) = onnx::Sub(%4006, %4007) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %4009 : Float(*, 384, *, *, device=cpu) = onnx::Mul(%4008, %4008) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %4010 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%4009) # /tmp/ipykernel_26455/952968418.py:32:0\n", " %4011 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::ReduceMean[axes=[1], keepdims=1](%4006) # /tmp/ipykernel_26455/952968418.py:33:0\n", " %4012 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sub(%4006, %4011) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %4013 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1e-05}]()\n", " %4014 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%4010, %4013) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %4015 : Float(*, 1, *, *, strides=[256, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Sqrt(%4014) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %4016 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Div(%4012, %4015) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %4017 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%4016, %layers.2.2.layers.9.1.norm.g) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %input.620 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%4017, %layers.2.2.layers.9.1.norm.b) # /tmp/ipykernel_26455/952968418.py:34:0\n", " %4019 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=0, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.620, %layers.2.2.layers.9.1.fn.net.0.weight, %layers.2.2.layers.9.1.fn.net.0.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %4020 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1.41421}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4021 : Float(*, 1536, *, *, device=cpu) = onnx::Div(%4019, %4020) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4022 : Float(*, 1536, *, *, device=cpu) = onnx::Erf(%4021) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4023 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4024 : Float(*, 1536, *, *, device=cpu) = onnx::Add(%4022, %4023) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4025 : Float(*, 1536, *, *, device=cpu) = onnx::Mul(%4019, %4024) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %4026 : Float(requires_grad=0, device=cpu) = onnx::Constant[value={0.5}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.624 : Float(*, 1536, *, *, strides=[393216, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Mul(%4025, %4026) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/activation.py:670:0\n", " %input.628 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%input.624, %layers.2.2.layers.9.1.fn.net.3.weight, %layers.2.2.layers.9.1.fn.net.3.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/conv.py:443:0\n", " %input.632 : Float(*, 384, *, *, strides=[98304, 256, 16, 1], requires_grad=1, device=cpu) = onnx::Add(%input.628, %4006) # /tmp/ipykernel_26455/952968418.py:115:0\n", " %tensor.519 : Float(*, 384, 1, 1, strides=[384, 1, 1, 1], requires_grad=1, device=cpu) = onnx::GlobalAveragePool(%input.632) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/functional.py:1241:0\n", " %4031 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.519) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4032 : Long(device=cpu) = onnx::Constant[value={0}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4033 : Long(device=cpu) = onnx::Gather[axis=0](%4031, %4032) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4034 : Long(4, strides=[1], device=cpu) = onnx::Shape(%tensor.519) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4035 : Long(device=cpu) = onnx::Constant[value={1}]() # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4036 : Long(device=cpu) = onnx::Gather[axis=0](%4034, %4035) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:66:0\n", " %4037 : Long(requires_grad=0, device=cpu) = onnx::Constant[value={1}]()\n", " %4038 : Long(requires_grad=0, device=cpu) = onnx::Mul(%4033, %4037) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %4039 : Long(requires_grad=0, device=cpu) = onnx::Mul(%4038, %4036) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/einops.py:36:0\n", " %4040 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %4041 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%4039, %4040)\n", " %4042 : Long(1, strides=[1], device=cpu) = onnx::Concat[axis=0](%4041) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:70:0\n", " %tensor.523 : Float(*, strides=[1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.519, %4042) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:70:0\n", " %4044 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %4045 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%4033, %4044)\n", " %4046 : Long(1, strides=[1], device=cpu) = onnx::Constant[value={0}]()\n", " %4047 : Long(1, strides=[1], device=cpu) = onnx::Unsqueeze(%4036, %4046)\n", " %4048 : Long(2, strides=[1], device=cpu) = onnx::Concat[axis=0](%4045, %4047) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:70:0\n", " %4049 : Float(*, 384, strides=[384, 1], requires_grad=1, device=cpu) = onnx::Reshape(%tensor.523, %4048) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/einops/_torch_specific.py:70:0\n", " %output : Float(*, 2, strides=[2, 1], requires_grad=1, device=cpu) = onnx::Gemm[alpha=1., beta=1., transB=1](%4049, %to_logits.2.weight, %to_logits.2.bias) # /home/gautham/anaconda3/envs/torch1.11/lib/python3.8/site-packages/torch/nn/modules/linear.py:103:0\n", " return (%output)\n", "\n" ] } ], "source": [ "onnx_path = '../model/model.onnx'\n", "\n", "torch.onnx.export(model,\n", " dummy_input,\n", " onnx_path,\n", " verbose=True,\n", " input_names = ['input'], # the model's input names\n", " output_names = ['output'], # the model's output names\n", " dynamic_axes={'input' : {0 : 'batch_size'}, # variable length axes\n", " 'output' : {0 : 'batch_size'}},\n", " opset_version=13)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Verifying the ONNX model" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "import onnx\n", "\n", "onnx_model = onnx.load(onnx_path)\n", "onnx.checker.check_model(onnx_model)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Comparing ONNX Runtime and PyTorch results" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Exported model has been tested with ONNXRuntime, and the result looks good!\n" ] } ], "source": [ "import onnxruntime\n", "import numpy as np\n", "\n", "ort_session = onnxruntime.InferenceSession(onnx_path)\n", "\n", "def to_numpy(tensor):\n", " return tensor.detach().cpu().numpy() if tensor.requires_grad else tensor.cpu().numpy()\n", "\n", "ort_inputs = {ort_session.get_inputs()[0].name: to_numpy(dummy_input)}\n", "ort_outs = ort_session.run(None, ort_inputs)\n", "\n", "np.testing.assert_allclose(to_numpy(torch_out), ort_outs[0], rtol=1e-03, atol=1e-05)\n", "\n", "print(\"Exported model has been tested with ONNXRuntime, and the result looks good!\")\n" ] } ], "metadata": { "kernelspec": { "display_name": "torch1.11", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.16" }, "orig_nbformat": 4, "vscode": { "interpreter": { "hash": "f23f8e2d4ed0a75a8c70e383ed5d654264680913d7fb3d20a1012009bf9528b0" } } }, "nbformat": 4, "nbformat_minor": 2 }