Reminder
Reproduction
CUDA_VISIBLE_DEVICES=9 python3.11 src/train.py
--stage rm
--do_train True
--model_name_or_path /data/hf_deploy/internlm2-7b-human-v2_merge_sft
--create_new_adapter
--dataset hh_rlhf_prompt
--template intern
--finetuning_type lora
--lora_target wqkv
--output_dir /data/LLaMA-Factory-main/results/intern_hh_rw
--per_device_train_batch_size 1
--gradient_accumulation_steps 4
--lr_scheduler_type cosine
--logging_steps 5
--save_steps 100
--learning_rate 1e-5
--num_train_epochs 2
--plot_loss
--overwrite_output_dir
--fp16
Expected behavior
rm 微调使用当前框架最新代码，报错提示没有语言模型头部,看了issue之前修复了chatglm2-6b的问题，使用最新框架训练internlm2还是会出现这个报错
System Info
Traceback (most recent call last):
File "/data/hf_deploy/LLaMA-Factory-main/src/train_bash.py", line 14, in
main()
File "/data/hf_deploy/LLaMA-Factory-main/src/train_bash.py", line 5, in main
run_exp()
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/train/tuner.py", line 33, in run_exp
run_rm(model_args, data_args, training_args, finetuning_args, callbacks)
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/train/rm/workflow.py", line 31, in run_rm
model, tokenizer = load_model_and_tokenizer(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/model/loader.py", line 195, in load_model_and_tokenizer
model: "AutoModelForCausalLMWithValueHead" = AutoModelForCausalLMWithValueHead.from_pretrained(model)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/miniconda3/lib/python3.11/site-packages/trl/models/modeling_base.py", line 276, in from_pretrained
model = cls(pretrained_model, **multi_adapter_args, **trl_model_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/miniconda3/lib/python3.11/site-packages/trl/models/modeling_value_head.py", line 113, in init
raise ValueError("The model does not have a language model head, please use a model that has one.")
ValueError: The model does not have a language model head, please use a model that has one.
Others
No response
Reminder
Reproduction
CUDA_VISIBLE_DEVICES=9 python3.11 src/train.py
--stage rm
--do_train True
--model_name_or_path /data/hf_deploy/internlm2-7b-human-v2_merge_sft
--create_new_adapter
--dataset hh_rlhf_prompt
--template intern
--finetuning_type lora
--lora_target wqkv
--output_dir /data/LLaMA-Factory-main/results/intern_hh_rw
--per_device_train_batch_size 1
--gradient_accumulation_steps 4
--lr_scheduler_type cosine
--logging_steps 5
--save_steps 100
--learning_rate 1e-5
--num_train_epochs 2
--plot_loss
--overwrite_output_dir
--fp16
Expected behavior
rm 微调使用当前框架最新代码，报错提示没有语言模型头部,看了issue之前修复了chatglm2-6b的问题，使用最新框架训练internlm2还是会出现这个报错
System Info
Traceback (most recent call last):
File "/data/hf_deploy/LLaMA-Factory-main/src/train_bash.py", line 14, in
main()
File "/data/hf_deploy/LLaMA-Factory-main/src/train_bash.py", line 5, in main
run_exp()
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/train/tuner.py", line 33, in run_exp
run_rm(model_args, data_args, training_args, finetuning_args, callbacks)
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/train/rm/workflow.py", line 31, in run_rm
model, tokenizer = load_model_and_tokenizer(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/hf_deploy/LLaMA-Factory-main/src/llmtuner/model/loader.py", line 195, in load_model_and_tokenizer
model: "AutoModelForCausalLMWithValueHead" = AutoModelForCausalLMWithValueHead.from_pretrained(model)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/miniconda3/lib/python3.11/site-packages/trl/models/modeling_base.py", line 276, in from_pretrained
model = cls(pretrained_model, **multi_adapter_args, **trl_model_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/miniconda3/lib/python3.11/site-packages/trl/models/modeling_value_head.py", line 113, in init
raise ValueError("The model does not have a language model head, please use a model that has one.")
ValueError: The model does not have a language model head, please use a model that has one.
Others
No response