Loading vagen/rollout/qwen_rollout/rollout_manager.py +3 −3 Original line number Original line Diff line number Diff line Loading @@ -39,13 +39,13 @@ class QwenVLRolloutManager(): 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True """ """ llm_raw_response = re.sub(r'<image>', '', llm_raw_response) llm_raw_response = llm_raw_response.replace('<image>', '') if prep_for_loss_mask: if prep_for_loss_mask: # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation sptk_b = self.config.special_token_for_loss_mask[0] sptk_b = self.config.special_token_for_loss_mask[0] sptk_e = self.config.special_token_for_loss_mask[1] sptk_e = self.config.special_token_for_loss_mask[1] llm_raw_response = re.sub(sptk_e, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_b, '') llm_raw_response = re.sub(sptk_b, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_e, '') llm_raw_response = sptk_b + llm_raw_response + sptk_e llm_raw_response = sptk_b + llm_raw_response + sptk_e return llm_raw_response return llm_raw_response Loading vagen/rollout/qwen_rollout/rollout_manager_service.py +3 −3 Original line number Original line Diff line number Diff line Loading @@ -43,13 +43,13 @@ class QwenVLRolloutManagerService(): 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True """ """ llm_raw_response = re.sub(r'<image>', '', llm_raw_response) llm_raw_response = llm_raw_response.replace('<image>', '') if prep_for_loss_mask: if prep_for_loss_mask: # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation sptk_b = self.config.special_token_for_loss_mask[0] sptk_b = self.config.special_token_for_loss_mask[0] sptk_e = self.config.special_token_for_loss_mask[1] sptk_e = self.config.special_token_for_loss_mask[1] llm_raw_response = re.sub(sptk_e, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_b, '') llm_raw_response = re.sub(sptk_b, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_e, '') llm_raw_response = sptk_b + llm_raw_response + sptk_e llm_raw_response = sptk_b + llm_raw_response + sptk_e return llm_raw_response return llm_raw_response Loading Loading
vagen/rollout/qwen_rollout/rollout_manager.py +3 −3 Original line number Original line Diff line number Diff line Loading @@ -39,13 +39,13 @@ class QwenVLRolloutManager(): 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True """ """ llm_raw_response = re.sub(r'<image>', '', llm_raw_response) llm_raw_response = llm_raw_response.replace('<image>', '') if prep_for_loss_mask: if prep_for_loss_mask: # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation sptk_b = self.config.special_token_for_loss_mask[0] sptk_b = self.config.special_token_for_loss_mask[0] sptk_e = self.config.special_token_for_loss_mask[1] sptk_e = self.config.special_token_for_loss_mask[1] llm_raw_response = re.sub(sptk_e, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_b, '') llm_raw_response = re.sub(sptk_b, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_e, '') llm_raw_response = sptk_b + llm_raw_response + sptk_e llm_raw_response = sptk_b + llm_raw_response + sptk_e return llm_raw_response return llm_raw_response Loading
vagen/rollout/qwen_rollout/rollout_manager_service.py +3 −3 Original line number Original line Diff line number Diff line Loading @@ -43,13 +43,13 @@ class QwenVLRolloutManagerService(): 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 1. Filter out special tokens: <image> and special tokens marking environment observation in the llm generated response 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True 2. prep_for_loss_mask: if true, add special tokens to the beginning and end of the response if compute_loss_mask is True """ """ llm_raw_response = re.sub(r'<image>', '', llm_raw_response) llm_raw_response = llm_raw_response.replace('<image>', '') if prep_for_loss_mask: if prep_for_loss_mask: # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation # filtering special tokens for llm_raw_response, then adding them to the beginning and end of the response for loss mask computation sptk_b = self.config.special_token_for_loss_mask[0] sptk_b = self.config.special_token_for_loss_mask[0] sptk_e = self.config.special_token_for_loss_mask[1] sptk_e = self.config.special_token_for_loss_mask[1] llm_raw_response = re.sub(sptk_e, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_b, '') llm_raw_response = re.sub(sptk_b, '', llm_raw_response) llm_raw_response = llm_raw_response.replace(sptk_e, '') llm_raw_response = sptk_b + llm_raw_response + sptk_e llm_raw_response = sptk_b + llm_raw_response + sptk_e return llm_raw_response return llm_raw_response Loading
