mirror of
https://github.com/hpcaitech/ColossalAI.git
synced 2025-07-18 17:31:53 +00:00
d20c8ffd97
* add grpo, support rlvr * add grpo, support rlvr * tested deepseek r1 pipeline * add ci * verify grpo r1 * verify grpo r1 * update readme, remove unused code * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * remove path * clean code * fix circular import * fix ci OOM * fix ci OOM * skip kto tp, fix qwen generation --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
48 lines
1.6 KiB
Python
Executable File
48 lines
1.6 KiB
Python
Executable File
"""
|
|
reward model
|
|
"""
|
|
|
|
from typing import Optional
|
|
|
|
import torch
|
|
import torch.nn as nn
|
|
from coati.models import BaseModel
|
|
from transformers import PretrainedConfig
|
|
|
|
|
|
class RewardModel(BaseModel):
|
|
"""
|
|
Reward model class.
|
|
|
|
Args:
|
|
pretrained str: huggingface or local model path
|
|
config: PretrainedConfig object
|
|
**kwargs: all other kwargs as in AutoModel.from_pretrained
|
|
"""
|
|
|
|
def __init__(self, pretrained: str = None, config: Optional[PretrainedConfig] = None, **kwargs) -> None:
|
|
super().__init__(pretrained=pretrained, config=config, **kwargs)
|
|
self.value_head = nn.Linear(self.last_hidden_state_size, 1)
|
|
self.value_head.weight.data.normal_(mean=0.0, std=1 / (self.last_hidden_state_size + 1))
|
|
|
|
def forward(
|
|
self, input_ids: torch.LongTensor, attention_mask: Optional[torch.Tensor] = None, **kwargs
|
|
) -> torch.Tensor:
|
|
outputs = self.model(input_ids, attention_mask=attention_mask)
|
|
|
|
last_hidden_states = outputs["last_hidden_state"]
|
|
sequence_lengths = torch.max(attention_mask * torch.arange(input_ids.size(1), device=input_ids.device), dim=1)[
|
|
0
|
|
]
|
|
sequence_hidden_states = last_hidden_states[torch.arange(last_hidden_states.size(0)), sequence_lengths].type(
|
|
self.value_head.weight.dtype
|
|
)
|
|
values = self.value_head(sequence_hidden_states).squeeze(-1) # Ensure shape is (B,)
|
|
return values
|
|
|
|
def get_input_embeddings(self):
|
|
return self.model.get_input_embeddings()
|
|
|
|
def get_output_embeddings(self):
|
|
return self.model.get_output_embeddings()
|