kyy/llm_trainer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

modify style

Browse Source
This commit is contained in:
BUAADreamer
2024-04-25 21:15:16 +08:00
parent 43d7ad5ecc
commit 1dcabafe72
16 changed files with 374 additions and 502 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
src/llmtuner/model/adapter.py
View File

@@ -38,9 +38,7 @@ def init_adapter(
logger.info("Adapter is not found at evaluation, load the base model.")
return model
if finetuning_args.finetuning_type != "lora" and getattr(
model, "quantization_method", None
):
if finetuning_args.finetuning_type != "lora" and getattr(model, "quantization_method", None):
raise ValueError("You can only use lora for quantized models.")
if finetuning_args.finetuning_type == "full" and is_trainable:
@@ -68,12 +66,8 @@ def init_adapter(
stride = num_layers // finetuning_args.num_layer_trainable
trainable_layer_ids = range(stride - 1, num_layers + stride - 1, stride)
elif (
finetuning_args.num_layer_trainable > 0
): # fine-tuning the last n layers if num_layer_trainable > 0
trainable_layer_ids = range(
num_layers - finetuning_args.num_layer_trainable, num_layers
)
elif finetuning_args.num_layer_trainable > 0: # fine-tuning the last n layers if num_layer_trainable > 0
trainable_layer_ids = range(num_layers - finetuning_args.num_layer_trainable, num_layers)
else: # fine-tuning the first n layers if num_layer_trainable < 0
trainable_layer_ids = range(-finetuning_args.num_layer_trainable)
@@ -88,15 +82,11 @@ def init_adapter(
for module_name in finetuning_args.name_module_trainable:
if module_name not in freeze_modules:
raise ValueError(
"Module {} is not found, please choose from {}".format(
module_name, ", ".join(freeze_modules)
)
"Module {} is not found, please choose from {}".format(module_name, ", ".join(freeze_modules))
)
for idx in trainable_layer_ids:
trainable_layers.append(
".{:d}.{}".format(idx, module_name if module_name != "all" else "")
)
trainable_layers.append(".{:d}.{}".format(idx, module_name if module_name != "all" else ""))
for name, param in model.named_parameters():
if any(trainable_layer in name for trainable_layer in trainable_layers):
@@ -105,43 +95,27 @@ def init_adapter(
else:
param.requires_grad_(False)
logger.info(
"Set trainable layers: {}".format(",".join(map(str, trainable_layer_ids)))
)
logger.info("Set trainable layers: {}".format(",".join(map(str, trainable_layer_ids))))
if finetuning_args.finetuning_type == "lora":
logger.info(
"Fine-tuning method: {}".format(
"DoRA" if finetuning_args.use_dora else "LoRA"
)
)
logger.info("Fine-tuning method: {}".format("DoRA" if finetuning_args.use_dora else "LoRA"))
adapter_to_resume = None
if model_args.adapter_name_or_path is not None:
is_mergeable = True
if getattr(
model, "quantization_method", None
): # merge lora in quantized model is unstable
assert (
len(model_args.adapter_name_or_path) == 1
), "Quantized model only accepts a single adapter."
if getattr(model, "quantization_method", None): # merge lora in quantized model is unstable
assert len(model_args.adapter_name_or_path) == 1, "Quantized model only accepts a single adapter."
is_mergeable = False
if is_deepspeed_zero3_enabled():
assert (
len(model_args.adapter_name_or_path) == 1
), "Cannot use multiple adapters in DeepSpeed ZeRO-3."
assert len(model_args.adapter_name_or_path) == 1, "Cannot use multiple adapters in DeepSpeed ZeRO-3."
is_mergeable = False
if model_args.use_unsloth:
assert (
len(model_args.adapter_name_or_path) == 1
), "Unsloth model only accepts a single adapter."
assert len(model_args.adapter_name_or_path) == 1, "Unsloth model only accepts a single adapter."
is_mergeable = False
if (is_trainable and not finetuning_args.create_new_adapter) or (
not is_mergeable
):
if (is_trainable and not finetuning_args.create_new_adapter) or (not is_mergeable):
adapter_to_merge = model_args.adapter_name_or_path[:-1]
adapter_to_resume = model_args.adapter_name_or_path[-1]
else:
@@ -158,9 +132,7 @@ def init_adapter(
if adapter_to_resume is not None: # resume lora training
if model_args.use_unsloth:
model = load_unsloth_peft_model(
config, model_args, is_trainable=is_trainable
)
model = load_unsloth_peft_model(config, model_args, is_trainable=is_trainable)
else:
model = PeftModel.from_pretrained(
model,
@@ -169,27 +141,19 @@ def init_adapter(
offload_folder=model_args.offload_folder,
)
if (
is_trainable and adapter_to_resume is None
): # create new lora weights while training
if (
len(finetuning_args.lora_target) == 1
and finetuning_args.lora_target[0] == "all"
):
if is_trainable and adapter_to_resume is None: # create new lora weights while training
if len(finetuning_args.lora_target) == 1 and finetuning_args.lora_target[0] == "all":
target_modules = find_all_linear_modules(model)
else:
target_modules = finetuning_args.lora_target
if finetuning_args.use_llama_pro:
target_modules = find_expanded_modules(
model, target_modules, finetuning_args.num_layer_trainable
)
target_modules = find_expanded_modules(model, target_modules, finetuning_args.num_layer_trainable)
if (
finetuning_args.use_dora
and getattr(model, "quantization_method", None) is not None
and getattr(model, "quantization_method", None)
!= QuantizationMethod.BITS_AND_BYTES
and getattr(model, "quantization_method", None) != QuantizationMethod.BITS_AND_BYTES
):
raise ValueError("DoRA is not compatible with PTQ-quantized models.")
@@ -202,11 +166,7 @@ def init_adapter(
module_names.add(name.split(".")[-1])
finetuning_args.additional_target = module_names
logger.warning(
"Vocab has been resized, add {} to trainable params.".format(
",".join(module_names)
)
)
logger.warning("Vocab has been resized, add {} to trainable params.".format(",".join(module_names)))
peft_kwargs = {
"r": finetuning_args.lora_rank,
@@ -233,10 +193,6 @@ def init_adapter(
param.data = param.data.to(torch.float32)
if model_args.adapter_name_or_path is not None:
logger.info(
"Loaded adapter(s): {}".format(
",".join(model_args.adapter_name_or_path)
)
)
logger.info("Loaded adapter(s): {}".format(",".join(model_args.adapter_name_or_path)))
return model