Spaces:
Running
on
Zero
Running
on
Zero
File size: 9,359 Bytes
9fa3d89 |
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 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 |
# Copyright 2023 Haotian Liu
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import List, Optional, Tuple, Union
import torch
import torch.nn as nn
from transformers import AutoConfig, AutoModelForCausalLM, \
Phi3Config, Phi3Model, Phi3ForCausalLM
from transformers.modeling_outputs import CausalLMOutputWithPast
from dataclasses import dataclass
from ..ola_arch import OlaLlavaMetaModel, OlaLlavaMetaForCausalLM
import torch.distributed as dist
try:
import wandb
except:
pass
import os
from torch.nn import CrossEntropyLoss
from .base_lm import BaseCausalLM
from .base_ola_vlm import BaseOLA_VLM
@dataclass
class OlaCausalLLMOutputWithPast(CausalLMOutputWithPast):
image_embs: Optional[Tuple[torch.FloatTensor]] = None
seg_embs: Optional[Tuple[torch.FloatTensor]] = None
depth_embs: Optional[Tuple[torch.FloatTensor]] = None
depth_preds: Optional[Tuple[torch.FloatTensor]] = None
class OlaLlavaPhi3Config(Phi3Config):
model_type = "ola_phi3"
class OlaLlavaPhi3Model(OlaLlavaMetaModel, Phi3Model):
config_class = OlaLlavaPhi3Config
def __init__(self, config: Phi3Config):
super(OlaLlavaPhi3Model, self).__init__(config)
class OlaLlavaPhi3ForCausalLM(Phi3ForCausalLM, OlaLlavaMetaForCausalLM, BaseOLA_VLM):
config_class = OlaLlavaPhi3Config
def __init__(self, config):
super(Phi3ForCausalLM, self).__init__(config)
self.model = OlaLlavaPhi3Model(config)
self.vocab_size = config.vocab_size
self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
self.config = config
self.is_use_reference_model = False
self.NUM_SYS_TOKENS = 13
# Initialize weights and apply final processing
self.post_init()
def get_model(self):
return self.model
def _forward(
self,
input_ids: torch.LongTensor = None,
attention_mask: Optional[torch.Tensor] = None,
position_ids: Optional[torch.LongTensor] = None,
past_key_values: Optional[List[torch.FloatTensor]] = None,
inputs_embeds: Optional[torch.FloatTensor] = None,
labels: Optional[torch.LongTensor] = None,
use_cache: Optional[bool] = None,
output_attentions: Optional[bool] = None,
output_hidden_states: Optional[bool] = None,
return_dict: Optional[bool] = None,
pil_images = None,
gen_mask: Optional[torch.FloatTensor] = None,
seg_mask: Optional[torch.FloatTensor] = None,
depth_mask: Optional[torch.FloatTensor] = None,
) -> Union[Tuple, OlaCausalLLMOutputWithPast]:
output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
output_hidden_states = (
output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
)
return_dict = return_dict if return_dict is not None else self.config.use_return_dict
# decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
outputs = self.model(
input_ids=input_ids,
attention_mask=attention_mask,
position_ids=position_ids,
past_key_values=past_key_values,
inputs_embeds=inputs_embeds,
use_cache=use_cache,
output_attentions=output_attentions,
output_hidden_states=True,
return_dict=return_dict,
)
hidden_states = outputs[0]
layer_states = outputs[-1][1:]
logits = self.lm_head(hidden_states)
logits = logits.float()
text_loss = None
loss = None
if labels is not None:
# Shift so that tokens < n predict n
shift_logits = logits[..., :-1, :].contiguous()
shift_labels = labels[..., 1:].contiguous()
# Flatten the tokens
loss_fct = CrossEntropyLoss()
shift_logits = shift_logits.view(-1, self.config.vocab_size)
shift_labels = shift_labels.view(-1)
# Enable model parallelism
shift_labels = shift_labels.to(shift_logits.device)
text_loss = loss_fct(shift_logits, shift_labels)
dinov2_emb, dinov2_loss = self.dinov2_emb_forward(pil_images, layer_states, hidden_states, seg_mask)
depth_preds, depth_embs, depth_loss, depth_l1_loss, depth_cont_loss = self.depth_emb_forward(pil_images, layer_states, depth_mask)
seg_embs, seg_loss, seg_l1_loss, seg_contrastive_loss = self.seg_emb_forward(pil_images, hidden_states, layer_states, seg_mask)
img_embs, gen_loss, gen_mse_loss, gen_con_loss = self.gen_emb_forward(pil_images, hidden_states, layer_states, gen_mask)
if text_loss is not None:
loss = text_loss + seg_loss + depth_loss + gen_loss + dinov2_loss
try:
if dist.get_rank() == 0:
if loss > text_loss:
log_dict = {
"depth_loss": depth_loss,
"gen_loss": gen_loss,
"depth_l1_loss": depth_l1_loss,
"depth_contrastive_loss": depth_cont_loss,
"dinov2_loss": dinov2_loss,
"gen_mse_loss": gen_mse_loss,
"gen_contrastive_loss": gen_con_loss,
"seg_loss": seg_loss,
"seg_l1_loss": seg_l1_loss,
"seg_cosine-emb_loss": seg_ce_loss,
"seg_contrastive_loss": seg_contrastive_loss,
"text_loss": text_loss,
"loss": loss,
}
filtered_log_dict = {key: value for key, value in log_dict.items() if value > 0}
wandb.log(filtered_log_dict)
else:
wandb.log({
"text_loss": text_loss,
"loss": loss,
})
self.steps += 1
except:
pass
if not return_dict:
output = (logits,) + outputs[1:]
return (loss,) + output if loss is not None else output
return OlaCausalLLMOutputWithPast(
loss=loss,
logits=logits,
past_key_values=outputs.past_key_values,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
image_embs=img_embs,
seg_embs=seg_embs,
depth_embs=depth_embs,
depth_preds=depth_preds,
)
def forward(
self,
input_ids: torch.LongTensor = None,
attention_mask: Optional[torch.Tensor] = None,
position_ids: Optional[torch.LongTensor] = None,
past_key_values: Optional[List[torch.FloatTensor]] = None,
inputs_embeds: Optional[torch.FloatTensor] = None,
labels: Optional[torch.LongTensor] = None,
use_cache: Optional[bool] = None,
output_attentions: Optional[bool] = None,
output_hidden_states: Optional[bool] = None,
images: Optional[torch.FloatTensor] = None,
image_sizes: Optional[List[List[int]]] = None,
return_dict: Optional[bool] = None,
pil_images: Optional[List[object]] = None,
gen_mask: Optional[torch.FloatTensor] = None,
seg_mask: Optional[torch.FloatTensor] = None,
depth_mask: Optional[torch.FloatTensor] = None,
**kwargs,
) -> Union[Tuple, CausalLMOutputWithPast]:
if inputs_embeds is None:
(
input_ids,
position_ids,
attention_mask,
past_key_values,
inputs_embeds,
labels,
) = self.prepare_inputs_labels_for_multimodal(
input_ids,
position_ids,
attention_mask,
past_key_values,
labels,
images,
image_sizes
)
return self._forward(
input_ids=input_ids,
attention_mask=attention_mask,
position_ids=position_ids,
past_key_values=past_key_values,
inputs_embeds=inputs_embeds,
labels=labels,
use_cache=use_cache,
output_attentions=output_attentions,
output_hidden_states=output_hidden_states,
return_dict=return_dict,
pil_images=pil_images,
gen_mask=gen_mask,
seg_mask=seg_mask,
depth_mask=depth_mask,
)
AutoConfig.register("ola_phi3", OlaLlavaPhi3Config)
AutoModelForCausalLM.register(OlaLlavaPhi3Config, OlaLlavaPhi3ForCausalLM) |