Hugging Face's logo

pxovela
/
cosmopolitan

TensorBoard
Model card Files Metrics Community
cosmopolitan
File size: 2,379 Bytes 
196ee88
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "# Load the checkpoint\n",
    "input_state_dict = torch.load(\"vodka_v5_4_768-ep60-gs146640.ckpt\")\n",
    "\n",
    "root_is_state_dict = False\n",
    "if \"state_dict\" in input_state_dict:\n",
    "    root_is_state_dict = True\n",
    "    state_dict = input_state_dict[\"state_dict\"]\n",
    "else:\n",
    "    state_dict = input_state_dict\n",
    "\n",
    "mappings = {\n",
    "    \"attn_1.to_out.0\": \"attn_1.proj_out\",\n",
    "    \"attn_1.to_k\": \"attn_1.k\",\n",
    "    \"attn_1.to_q\": \"attn_1.q\",\n",
    "    \"attn_1.to_v\": \"attn_1.v\"\n",
    "}\n",
    "\n",
    "def replace_all(key):\n",
    "    for mapping in mappings:\n",
    "        key = key.replace(mapping, mappings[mapping])\n",
    "    return key\n",
    "\n",
    "# First, create a new state_dict with renamed keys\n",
    "renamed_state_dict = dict()\n",
    "for key in state_dict:\n",
    "    renamed_state_dict[replace_all(key)] = state_dict[key]\n",
    "\n",
    "# Then, reshape the tensors in the renamed state_dict\n",
    "def reshape_tensors(sd):\n",
    "    for key in sd:\n",
    "        if \"attn_1\" in key and sd[key].shape == torch.Size([512, 512]):\n",
    "            sd[key] = sd[key].unsqueeze(2).unsqueeze(3)\n",
    "    return sd\n",
    "\n",
    "output_state_dict = reshape_tensors(renamed_state_dict)\n",
    "\n",
    "# Finally, save the updated state_dict\n",
    "if root_is_state_dict:\n",
    "    input_state_dict[\"state_dict\"] = output_state_dict\n",
    "    torch.save(input_state_dict, \"6_vodka_v5_768_adamw8bit_ep60.ckpt\")\n",
    "else:\n",
    "    torch.save(output_state_dict, \"6_vodka_v5_768_adamw8bit_ep60.ckpt\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "fastai",
   "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": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}