add llava and instructblip
Former-commit-id: 142fb6f4541a1acfefe66ff2574dabde53b00c06
This commit is contained in:
BUAADreamer
@@ -5,7 +5,7 @@ from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
|
||||
|
||||
import numpy as np
|
||||
import torch
|
||||
from transformers import Seq2SeqTrainer
|
||||
from transformers import Seq2SeqTrainer, Trainer
|
||||
|
||||
from ...extras.constants import IGNORE_INDEX
|
||||
from ...extras.logging import get_logger
|
||||
@@ -32,23 +32,6 @@ class CustomSeq2SeqTrainer(Seq2SeqTrainer):
|
||||
|
||||
self.accelerator.clip_grad_norm_ = MethodType(clip_grad_norm_for_sparse_tensor, self.accelerator)
|
||||
|
||||
# def compute_loss(self, model, inputs, return_outputs=False):
|
||||
# print(inputs.keys())
|
||||
# device = "cuda"
|
||||
# input_ids = inputs.get("input_ids").to(device)
|
||||
# pixel_values = inputs.get("pixel_values").to(device, torch.float16)
|
||||
# attention_mask = inputs.get("attention_mask").to(device)
|
||||
# labels = inputs.get("labels").to(device)
|
||||
#
|
||||
# outputs = model(input_ids=input_ids,
|
||||
# pixel_values=pixel_values,
|
||||
# labels=labels,
|
||||
# # attention_mask=attention_mask,
|
||||
# )
|
||||
# loss = outputs.loss
|
||||
# print("Loss:", loss.item())
|
||||
# return (loss, outputs) if return_outputs else loss
|
||||
|
||||
def create_optimizer(self) -> "torch.optim.Optimizer":
|
||||
if self.optimizer is None:
|
||||
self.optimizer = create_custom_optimzer(self.model, self.args, self.finetuning_args)
|
||||
@@ -59,79 +42,3 @@ class CustomSeq2SeqTrainer(Seq2SeqTrainer):
|
||||
) -> "torch.optim.lr_scheduler.LRScheduler":
|
||||
create_custom_scheduler(self.args, num_training_steps, optimizer)
|
||||
return super().create_scheduler(num_training_steps, optimizer)
|
||||
|
||||
def prediction_step(
|
||||
self,
|
||||
model: "torch.nn.Module",
|
||||
inputs: Dict[str, Union[torch.Tensor, Any]],
|
||||
prediction_loss_only: bool,
|
||||
ignore_keys: Optional[List[str]] = None,
|
||||
) -> Tuple[Optional[float], Optional[torch.Tensor], Optional[torch.Tensor]]:
|
||||
r"""
|
||||
Removes the prompt part in the generated tokens.
|
||||
|
||||
Subclass and override to inject custom behavior.
|
||||
"""
|
||||
labels = inputs["labels"].detach().clone() if "labels" in inputs else None # backup labels
|
||||
if self.args.predict_with_generate:
|
||||
assert self.tokenizer.padding_side == "left", "This method only accepts left-padded tensor."
|
||||
prompt_len, label_len = inputs["input_ids"].size(-1), inputs["labels"].size(-1)
|
||||
if prompt_len > label_len:
|
||||
inputs["labels"] = self._pad_tensors_to_target_len(inputs["labels"], inputs["input_ids"])
|
||||
if label_len > prompt_len: # truncate the labels instead of padding the inputs (llama2 fp16 compatibility)
|
||||
inputs["labels"] = inputs["labels"][:, :prompt_len]
|
||||
|
||||
loss, generated_tokens, _ = super().prediction_step( # ignore the returned labels (may be truncated)
|
||||
model, inputs, prediction_loss_only=prediction_loss_only, ignore_keys=ignore_keys
|
||||
)
|
||||
if generated_tokens is not None and self.args.predict_with_generate:
|
||||
generated_tokens[:, :prompt_len] = self.tokenizer.pad_token_id
|
||||
generated_tokens = generated_tokens.contiguous()
|
||||
|
||||
return loss, generated_tokens, labels
|
||||
|
||||
def _pad_tensors_to_target_len(self, src_tensor: torch.Tensor, tgt_tensor: torch.Tensor) -> torch.Tensor:
|
||||
r"""
|
||||
Pads the tensor to the same length as the target tensor.
|
||||
"""
|
||||
assert self.tokenizer.pad_token_id is not None, "Pad token is required."
|
||||
padded_tensor = self.tokenizer.pad_token_id * torch.ones_like(tgt_tensor)
|
||||
padded_tensor[:, -src_tensor.shape[-1]:] = src_tensor # adopt left-padding
|
||||
return padded_tensor.contiguous() # in contiguous memory
|
||||
|
||||
def save_predictions(self, predict_results: "PredictionOutput") -> None:
|
||||
r"""
|
||||
Saves model predictions to `output_dir`.
|
||||
|
||||
A custom behavior that not contained in Seq2SeqTrainer.
|
||||
"""
|
||||
if not self.is_world_process_zero():
|
||||
return
|
||||
|
||||
output_prediction_file = os.path.join(self.args.output_dir, "generated_predictions.jsonl")
|
||||
logger.info(f"Saving prediction results to {output_prediction_file}")
|
||||
|
||||
labels = np.where(
|
||||
predict_results.label_ids != IGNORE_INDEX, predict_results.label_ids, self.tokenizer.pad_token_id
|
||||
)
|
||||
preds = np.where(
|
||||
predict_results.predictions != IGNORE_INDEX, predict_results.predictions, self.tokenizer.pad_token_id
|
||||
)
|
||||
|
||||
for i in range(len(preds)):
|
||||
pad_len = np.nonzero(preds[i] != self.tokenizer.pad_token_id)[0]
|
||||
if len(pad_len):
|
||||
preds[i] = np.concatenate(
|
||||
(preds[i][pad_len[0]:], preds[i][: pad_len[0]]), axis=-1
|
||||
) # move pad token to last
|
||||
|
||||
decoded_labels = self.tokenizer.batch_decode(
|
||||
labels, skip_special_tokens=True, clean_up_tokenization_spaces=False
|
||||
)
|
||||
decoded_preds = self.tokenizer.batch_decode(preds, skip_special_tokens=True, clean_up_tokenization_spaces=True)
|
||||
|
||||
with open(output_prediction_file, "w", encoding="utf-8") as writer:
|
||||
res: List[str] = []
|
||||
for label, pred in zip(decoded_labels, decoded_preds):
|
||||
res.append(json.dumps({"label": label, "predict": pred}, ensure_ascii=False))
|
||||
writer.write("\n".join(res))
|
||||
|
||||
Reference in New Issue
Block a user