提交 5f46304e 编写于 作者: U u010280923

opt reward model

上级 6e8dbe4b
...@@ -69,6 +69,11 @@ class RewardModel(pl.LightningModule): ...@@ -69,6 +69,11 @@ class RewardModel(pl.LightningModule):
self.prompt_embed = nn.Parameter(torch.zeros(1, 1, dim)) self.prompt_embed = nn.Parameter(torch.zeros(1, 1, dim))
self.response_embed = nn.Parameter(torch.zeros(1, 1, dim)) self.response_embed = nn.Parameter(torch.zeros(1, 1, dim))
self.padding_embed = nn.Parameter(torch.zeros(1, 1, dim), requires_grad=False) self.padding_embed = nn.Parameter(torch.zeros(1, 1, dim), requires_grad=False)
self.prompt_response_mask_embed = torch.stack([
self.prompt_embed,
self.response_embed,
self.padding_embed
])
# reward 得分计算 # reward 得分计算
self.pred_reward = nn.Sequential( self.pred_reward = nn.Sequential(
...@@ -161,11 +166,7 @@ class RewardModel(pl.LightningModule): ...@@ -161,11 +166,7 @@ class RewardModel(pl.LightningModule):
# 如果为 True,则从 prompt_embed 中选,否则从 response_embed 中选 # 如果为 True,则从 prompt_embed 中选,否则从 response_embed 中选
extra_embed = None extra_embed = None
if exists(prompt_mask): if exists(prompt_mask):
extra_embed = torch.where( extra_embed = self.prompt_response_mask_embed[prompt_mask]
rearrange(prompt_mask, 'b n -> b n 1'),
self.prompt_embed,
self.response_embed
)
# 获得最后一个 token 的 embedding # 获得最后一个 token 的 embedding
last_token_embeds = self.rwkv( last_token_embeds = self.rwkv(
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册