diff --git a/.gitattributes b/.gitattributes index cea30e3c58f9c9e909f9928ed4d095f2988157f8..e1c9a61388384b329c8814a46f54d02b2015acd0 100644 --- a/.gitattributes +++ b/.gitattributes @@ -65,3 +65,10 @@ old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/global_step_328/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text old_template_backup/precise_coder/qwen35-4b-GRPO_U/global_step_200/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-rej32/global_step_32/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-dsv4pro-1ep/global_step_41/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-dsv4pro-hard-2ep/global_step_21/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +precise_coder_sft/4b-SFT-glm-1ep/global_step_31/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/best_ckpt_state.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..6eb08094720ec594f3f2a69ee83c4d57c6f22dc6 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.5869523406801406, "best_global_step": 328, "no_improve_rounds": 0} \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/fsdp_config.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/config.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e8cd2d6f8b5c82e83b6479da2639aa4f7f24473 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": true, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/global_step_328/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/4b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/fsdp_config.json b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/config.json b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e8cd2d6f8b5c82e83b6479da2639aa4f7f24473 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": true, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/global_step_300/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/4b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/best_ckpt_state.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3121da8319c37e6e4ae570ea4a83f34ae19c4e28 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.6613064573701701, "best_global_step": 250, "no_improve_rounds": 2} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/fsdp_config.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/config.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..9767c2df7a0eb06a928081cfe6610a684c2ba51d --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": false, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": false, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 27, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1152, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4304, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 4096, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-27b-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/best_ckpt_state.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..9ffee76bfdbef8d078aabe94e5c16a5e07467d78 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.6441647629790495, "best_global_step": 250, "no_improve_rounds": 1} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/fsdp_config.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/config.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..9767c2df7a0eb06a928081cfe6610a684c2ba51d --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": false, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": false, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 27, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1152, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4304, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 4096, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/global_step_250/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..f1efb2054088985fad3cec9900b3f4a7561597b3 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-dsv4pro-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +300 \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/best_ckpt_state.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..89b72e56ae1ecb8fc6e775f92bae3e93ef8a5ae3 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.6366156602267299, "best_global_step": 150, "no_improve_rounds": 4} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/fsdp_config.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/config.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..9767c2df7a0eb06a928081cfe6610a684c2ba51d --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": false, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": false, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 27, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1152, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4304, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 4096, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/global_step_150/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/9b-SFT-glm-rl_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e8cd2d6f8b5c82e83b6479da2639aa4f7f24473 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": true, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..e8930b6df94cf1501f2663918ee3ceb4d40c709a --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +364 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-GRPO_U/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-4b-GRPO_U/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..8a32cf7839cd770c9bf35526cff2a66a770e0951 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-GRPO_U/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +250 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/best_ckpt_state.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..b217d9e716276815ac75109394e11207946e2c68 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.5682609959180753, "best_global_step": 150, "no_improve_rounds": 4} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/fsdp_config.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/config.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e8cd2d6f8b5c82e83b6479da2639aa4f7f24473 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": true, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/global_step_150/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..fa8f08cb6ff8a30dd34a3b442009bdfc798a7316 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +150 diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/config.json b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a50c6c903b4ceb94eb74663bc79d91145c64000b --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 2560, + "initializer_range": 0.02, + "intermediate_size": 9216, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 1, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": true, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": true, + "transformers_version": "5.7.0", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 24, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1024, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4096, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 2560, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..cf71f578cf0d510a8e3062e26a6c05ed1d156c46 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.7.0", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/global_step_100/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..4701cc793111aafe3a1e6f48d82e5b8c8a38943e --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-4b-sftrl-rl_0.5f1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +150 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/fsdp_config.json b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..a585dec894e63da457d9440ec6aa7caa16d20860 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/chat_template.jinja @@ -0,0 +1,154 @@ +{%- set image_count = namespace(value=0) %} +{%- set video_count = namespace(value=0) %} +{%- macro render_content(content, do_vision_count, is_system_content=false) %} + {%- if content is string %} + {{- content }} + {%- elif content is iterable and content is not mapping %} + {%- for item in content %} + {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain images.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set image_count.value = image_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Picture ' ~ image_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|image_pad|><|vision_end|>' }} + {%- elif 'video' in item or item.type == 'video' %} + {%- if is_system_content %} + {{- raise_exception('System message cannot contain videos.') }} + {%- endif %} + {%- if do_vision_count %} + {%- set video_count.value = video_count.value + 1 %} + {%- endif %} + {%- if add_vision_id %} + {{- 'Video ' ~ video_count.value ~ ': ' }} + {%- endif %} + {{- '<|vision_start|><|video_pad|><|vision_end|>' }} + {%- elif 'text' in item %} + {{- item.text }} + {%- else %} + {{- raise_exception('Unexpected item type in content.') }} + {%- endif %} + {%- endfor %} + {%- elif content is none or content is undefined %} + {{- '' }} + {%- else %} + {{- raise_exception('Unexpected content type.') }} + {%- endif %} +{%- endmacro %} +{%- if not messages %} + {{- raise_exception('No messages provided.') }} +{%- endif %} +{%- if tools and tools is iterable and tools is not mapping %} + {{- '<|im_start|>system\n' }} + {{- "# Tools\n\nYou have access to the following functions:\n\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n" }} + {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {%- if content %} + {{- '\n\n' + content }} + {%- endif %} + {%- endif %} + {{- '<|im_end|>\n' }} +{%- else %} + {%- if messages[0].role == 'system' %} + {%- set content = render_content(messages[0].content, false, true)|trim %} + {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" %} + {%- set content = render_content(message.content, false)|trim %} + {%- if not(content.startswith('') and content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if ns.multi_step_tool %} + {{- raise_exception('No user query found in messages.') }} +{%- endif %} +{%- for message in messages %} + {%- set content = render_content(message.content, true)|trim %} + {%- if message.role == "system" %} + {%- if not loop.first %} + {{- raise_exception('System message must be at the beginning.') }} + {%- endif %} + {%- elif message.role == "user" %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- set reasoning_content = reasoning_content|trim %} + {%- if loop.index0 > ns.last_query_index %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {%- if loop.first %} + {%- if content|trim %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n\n' }} + {%- endif %} + {%- else %} + {{- '\n\n\n' }} + {%- endif %} + {%- if tool_call.arguments is defined %} + {%- for args_name, args_value in tool_call.arguments|items %} + {{- '\n' }} + {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} + {{- args_value }} + {{- '\n\n' }} + {%- endfor %} + {%- endif %} + {{- '\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.previtem and loop.previtem.role != "tool" %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if not loop.last and loop.nextitem.role != "tool" %} + {{- '<|im_end|>\n' }} + {%- elif loop.last %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- else %} + {{- raise_exception('Unexpected message role.') }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- else %} + {{- '\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..9767c2df7a0eb06a928081cfe6610a684c2ba51d --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": false, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": false, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 27, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1152, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4304, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 4096, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8f50cfe9f041954a2175966a9ef6f88480f4769 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "eos_token_id": 248046, + "output_attentions": false, + "output_hidden_states": false, + "pad_token_id": 248044, + "transformers_version": "5.10.2", + "use_cache": true +} diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..1d134cd298be1e3be25db393d93a1cefe80e3214 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/global_step_300/actor/huggingface/tokenizer_config.json @@ -0,0 +1,33 @@ +{ + "add_prefix_space": false, + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "image_token": "<|image_pad|>", + "is_local": true, + "local_files_only": false, + "model_max_length": 262144, + "model_specific_special_tokens": { + "audio_bos_token": "<|audio_start|>", + "audio_eos_token": "<|audio_end|>", + "audio_token": "<|audio_pad|>", + "image_token": "<|image_pad|>", + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" + }, + "pad_token": "<|endoftext|>", + "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", + "processor_class": "Qwen3VLProcessor", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null, + "video_token": "<|video_pad|>", + "vision_bos_token": "<|vision_start|>", + "vision_eos_token": "<|vision_end|>" +} diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..e8930b6df94cf1501f2663918ee3ceb4d40c709a --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_0.5F1_min/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +364 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-GRPO_U/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-9b-GRPO_U/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..8a32cf7839cd770c9bf35526cff2a66a770e0951 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-GRPO_U/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +250 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/best_ckpt_state.json b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..7e6687016573cf3bf6f0e494968932249af41d36 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.5803733236679957, "best_global_step": 200, "no_improve_rounds": 3} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/config.json b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/config.json new file mode 100644 index 0000000000000000000000000000000000000000..9767c2df7a0eb06a928081cfe6610a684c2ba51d --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/config.json @@ -0,0 +1,112 @@ +{ + "architectures": [ + "Qwen3_5ForConditionalGeneration" + ], + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248046, + "image_token_id": 248056, + "model_type": "qwen3_5", + "pad_token_id": 248044, + "text_config": { + "attention_bias": false, + "attention_dropout": 0.0, + "attn_output_gate": true, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 248044, + "full_attention_interval": 4, + "head_dim": 256, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention", + "linear_attention", + "linear_attention", + "linear_attention", + "full_attention" + ], + "linear_conv_kernel_dim": 4, + "linear_key_head_dim": 128, + "linear_num_key_heads": 16, + "linear_num_value_heads": 32, + "linear_value_head_dim": 128, + "mamba_ssm_dtype": "float32", + "max_position_embeddings": 262144, + "mlp_only_layers": [], + "model_type": "qwen3_5_text", + "mtp_num_hidden_layers": 0, + "mtp_use_dedicated_embeddings": false, + "num_attention_heads": 16, + "num_hidden_layers": 32, + "num_key_value_heads": 4, + "pad_token_id": null, + "partial_rotary_factor": 0.25, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "mrope_interleaved": true, + "mrope_section": [ + 11, + 11, + 10 + ], + "partial_rotary_factor": 0.25, + "rope_theta": 10000000, + "rope_type": "default" + }, + "tie_word_embeddings": false, + "use_cache": true, + "vocab_size": 248320 + }, + "tie_word_embeddings": false, + "transformers_version": "5.10.2", + "video_token_id": 248057, + "vision_config": { + "deepstack_visual_indexes": [], + "depth": 27, + "dtype": "bfloat16", + "hidden_act": "gelu_pytorch_tanh", + "hidden_size": 1152, + "in_channels": 3, + "initializer_range": 0.02, + "intermediate_size": 4304, + "model_type": "qwen3_5_vision", + "num_heads": 16, + "num_position_embeddings": 2304, + "out_hidden_size": 4096, + "patch_size": 16, + "spatial_merge_size": 2, + "temporal_patch_size": 2 + }, + "vision_end_token_id": 248054, + "vision_start_token_id": 248053 +} diff --git a/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/processor_config.json b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/global_step_200/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-eagrpo_a0.8_b0.05/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/best_ckpt_state.json b/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3a54f65cfc68ccf1b71be43efabb7694e44ac36a --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.6551225683059307, "best_global_step": 350, "no_improve_rounds": 0} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..0dbf139f8d1a85546880d56c8ff838638c454501 --- /dev/null +++ b/old_template_backup/precise_coder/qwen35-9b-rl-strreplace/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +350 \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/best_ckpt_state.json b/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/best_ckpt_state.json new file mode 100644 index 0000000000000000000000000000000000000000..75d45b4db8cc8da53bdf565f885afa66b630a357 --- /dev/null +++ b/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/best_ckpt_state.json @@ -0,0 +1 @@ +{"best_val_score": 0.8006791570040722, "best_global_step": 328, "no_improve_rounds": 0} \ No newline at end of file diff --git a/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/latest_checkpointed_iteration.txt b/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/latest_checkpointed_iteration.txt new file mode 100644 index 0000000000000000000000000000000000000000..41a4a819ada66b2e8f7fea82f094eb1314f2e9ca --- /dev/null +++ b/old_template_backup/precise_coder/qwen36-27b-GRPO_0.5F1/latest_checkpointed_iteration.txt @@ -0,0 +1 @@ +328 \ No newline at end of file diff --git a/precise_coder/4b_rl_bcbfix_eagrpo/global_step_100/data.pt b/precise_coder/4b_rl_bcbfix_eagrpo/global_step_100/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..bd9214536c5699138b9ecc037f0ea7a49ab23f56 --- /dev/null +++ b/precise_coder/4b_rl_bcbfix_eagrpo/global_step_100/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ddc8573db7b91f0df155e47a4a89f8842109374bacf10a4a01dcd23a3e919f5 +size 7316 diff --git a/precise_coder/4b_rl_bcbfix_k0.05_f1ur/global_step_325/data.pt b/precise_coder/4b_rl_bcbfix_k0.05_f1ur/global_step_325/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..5ce6f265692749db1c0c882782c1a1faed10f0fa --- /dev/null +++ b/precise_coder/4b_rl_bcbfix_k0.05_f1ur/global_step_325/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:515d8b6bd43059ddbf11c623a8a2bc138ceccde1b8132a8585c80184a86f1bda +size 7316 diff --git a/precise_coder/4b_rl_bcbfix_ugr/global_step_328/data.pt b/precise_coder/4b_rl_bcbfix_ugr/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..ade2e1340d6420c3745861a06e58f9c51f3f4bad --- /dev/null +++ b/precise_coder/4b_rl_bcbfix_ugr/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ec69718d494ffac3dcc14534a773bbe365d4e7d9b345ed0b51ecefa73b28d6e +size 7316 diff --git a/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_0.pt b/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..cdcd708b54268fae8ed7e7ed80f2751df82c2be3 --- /dev/null +++ b/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5d571331c0af9671ead57885adfe038dab98e81f26b4fa727e4f40e43869063 +size 14973 diff --git a/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..948bb6940f392759b6a38ecc33fb01c3141f6644 --- /dev/null +++ b/precise_coder/4b_rl_k0.0/global_step_325/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a66485746c14ee69156c4cb302242d00fb605cbe80a01b658085634ba373e12d +size 14909 diff --git a/precise_coder/4b_rl_k0.0/global_step_325/data.pt b/precise_coder/4b_rl_k0.0/global_step_325/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..8fe6730adb60fe2fd1055fa37fb46035560c2754 --- /dev/null +++ b/precise_coder/4b_rl_k0.0/global_step_325/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c7a0e28d8d2debd2682c7bc8f71773b09088a3fa7d99b68f7e1d65e792c1481 +size 7316 diff --git a/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_0.pt b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..fcce353c7d117339b4dd3ea0de59a35acfb4b178 --- /dev/null +++ b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fd6750a2669f8eeee4ed152012145367402f33d976c7973600d8a9b636db1223 +size 14973 diff --git a/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_1.pt b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..a852cd601489c5d29923218ab59fe6e0cb4a3e0c --- /dev/null +++ b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:70b24313780fd13268b33ed2866939d74eba30208e20ccc5648b158276486b0e +size 14973 diff --git a/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..5826f60133f943bf1f2d20532ff91ee91b21af8b --- /dev/null +++ b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:94437aa403944986338c7cec6b34b0f3ba00a06f697027059021a6903f75450c +size 14973 diff --git a/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_3.pt b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..c4c99c7196ccb1b3fd519eea9ba54c7cfc5cdf36 --- /dev/null +++ b/precise_coder/4b_rl_k0.33/global_step_328/actor/extra_state_world_size_4_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f1fdd1017a019b5d3b9bf891cb8f66a1e8c64831245166ef5b682a5575820860 +size 14909 diff --git a/precise_coder/4b_rl_k0.33/global_step_328/data.pt b/precise_coder/4b_rl_k0.33/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..d86615ee7c9e43b7c521bf25f4d128123944cc07 --- /dev/null +++ b/precise_coder/4b_rl_k0.33/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3851bd9baa4d93ce02fd4563e6180030bb342332cb0fa0c04612d2a8a68ad70d +size 7316 diff --git a/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_0.pt b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..a4a562222dc9c2e29c269a6be430e86194c116df --- /dev/null +++ b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d87c1a4e1e5ad7a97bb6b7aa9f9adf09ce5f14efe827fbab0775752585ba829b +size 14909 diff --git a/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..efd54185aaad85874a9be1ae9292fdc6990f0038 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1a05527bf279ddab95d99289f13fc5c13d989ce84cbf299cb15927d8e77fac4 +size 14909 diff --git a/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_3.pt b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..3e9c78c5bdc00c1cd803d231ec14be9d8b3847e1 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/actor/extra_state_world_size_4_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:379973ee7c5595cf5ce9ccf5f4feb499673c6a4adbbd5a75e71b0b37106770ad +size 14909 diff --git a/precise_coder/4b_rl_k0.33_f1ur/global_step_328/data.pt b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..3b5811cae37733f3033d859944c56309936e9739 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_f1ur/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d768123a3fae8309db0a7a3d8f35a585bf7f74512c4b30b6523956d992057306 +size 7316 diff --git a/precise_coder/4b_rl_k0.33_f1ur_think8k/global_step_328/data.pt b/precise_coder/4b_rl_k0.33_f1ur_think8k/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..2af6a126c13e9cda256cd022a42b07de25a5b021 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_f1ur_think8k/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:329412932ab142035735f17a6e23a48c109e721827e3188e88cfd04cbc61a49a +size 7316 diff --git a/precise_coder/4b_rl_k0.33_gated_v2/global_step_275/actor/huggingface/processor_config.json b/precise_coder/4b_rl_k0.33_gated_v2/global_step_275/actor/huggingface/processor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..33818c7f9e991ad735fd240209f4fa73e6c28c50 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_gated_v2/global_step_275/actor/huggingface/processor_config.json @@ -0,0 +1,60 @@ +{ + "image_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_processor_type": "Qwen2VLImageProcessor", + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "merge_size": 2, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "size": { + "longest_edge": 16777216, + "shortest_edge": 65536 + }, + "temporal_patch_size": 2 + }, + "processor_class": "Qwen3VLProcessor", + "video_processor": { + "do_convert_rgb": true, + "do_normalize": true, + "do_rescale": true, + "do_resize": true, + "do_sample_frames": true, + "fps": 2, + "image_mean": [ + 0.5, + 0.5, + 0.5 + ], + "image_std": [ + 0.5, + 0.5, + 0.5 + ], + "max_frames": 768, + "merge_size": 2, + "min_frames": 4, + "patch_size": 16, + "resample": 3, + "rescale_factor": 0.00392156862745098, + "return_metadata": false, + "size": { + "longest_edge": 25165824, + "shortest_edge": 4096 + }, + "temporal_patch_size": 2, + "video_processor_type": "Qwen3VLVideoProcessor" + } +} diff --git a/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_1.pt b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..88144c5f13f21363800c49302c1e2b9f92a1d5cc --- /dev/null +++ b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d01147bfbb0ef8ee55eb4d1d22418d7745f47528abd8eed3dfddcbb64505bd9 +size 14909 diff --git a/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..48fcd1fa7f2cbacfc826bf10bbb9852c8931a1aa --- /dev/null +++ b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:673d43b2df2948de88b8c9f4189c676e5a9fc33ea2c875c1513361c3e385c063 +size 14909 diff --git a/precise_coder/4b_rl_k0.33_gdpo/global_step_328/data.pt b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..3bcac67af4b55c3c327191ea1fb31a8a5e442c84 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_gdpo/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:836317daa63470e72e54ea823b75f4c95b65ff3fa9a5455a6662c205994d05ec +size 7316 diff --git a/precise_coder/4b_rl_k0.33_rec/global_step_250/actor/extra_state_world_size_4_rank_1.pt b/precise_coder/4b_rl_k0.33_rec/global_step_250/actor/extra_state_world_size_4_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..b6460735d7cb9fc029fdef1767baa88bcdc14edc --- /dev/null +++ b/precise_coder/4b_rl_k0.33_rec/global_step_250/actor/extra_state_world_size_4_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:479647c7633eaa7f10852498848d2b419f0446d95b84ca3cff821e5c8215cf3c +size 14973 diff --git a/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_0.pt b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..5c1146f370f5d3eb757d8d4f504f4ecf005c63f5 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:05fff5fd46a0e98ce13ee4cb728faa222035ce2b7d8e1383c8b668a5ff551865 +size 14973 diff --git a/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_1.pt b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..e03904c6000397401ef63d2951a0fee91efe219c --- /dev/null +++ b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ed1e6190ef53c08c61fa36e073a5da7e7be681ad9590801d285397d80ecf5a4 +size 14973 diff --git a/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..60031ed25856b440d9015549550c9a55efa0fd04 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20d6219f73f796761a2269fa9a9972252a7936c372041f4e44ef5e22376f083c +size 14973 diff --git a/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_3.pt b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..0faca6b1cba77006ec500efcd95066da5d6c90ff --- /dev/null +++ b/precise_coder/4b_rl_k0.33_run2/global_step_328/actor/extra_state_world_size_4_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:67d295cd1ff0957592ea4d74256564317d844dde14c0566724cd2dcddb4a089b +size 14973 diff --git a/precise_coder/4b_rl_k0.33_run2/global_step_328/data.pt b/precise_coder/4b_rl_k0.33_run2/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..85b456554d878c7c4b4c0c638116543640961bc9 --- /dev/null +++ b/precise_coder/4b_rl_k0.33_run2/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:441c28e1831e3131077594ea63ee723f1b6ecc80e8f053b6d3a120f4a6827c36 +size 7316 diff --git a/precise_coder/4b_rl_k0.67/global_step_325/actor/fsdp_config.json b/precise_coder/4b_rl_k0.67/global_step_325/actor/fsdp_config.json new file mode 100644 index 0000000000000000000000000000000000000000..17ff55630675ec23d6a1d79ec7a480aa25faa5de --- /dev/null +++ b/precise_coder/4b_rl_k0.67/global_step_325/actor/fsdp_config.json @@ -0,0 +1,4 @@ +{ + "FSDP_version": 2, + "world_size": 4 +} \ No newline at end of file diff --git a/precise_coder/4b_rl_k0.83/global_step_275/actor/extra_state_world_size_4_rank_3.pt b/precise_coder/4b_rl_k0.83/global_step_275/actor/extra_state_world_size_4_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..56bc7853480525275acc2082d101df95fa62e17a --- /dev/null +++ b/precise_coder/4b_rl_k0.83/global_step_275/actor/extra_state_world_size_4_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6f1719327745681a9511ccb96b8b22ad2fff21d72b113b985c439058ad748761 +size 15037 diff --git a/precise_coder/4b_rl_k1.0_f1ur/global_step_328/data.pt b/precise_coder/4b_rl_k1.0_f1ur/global_step_328/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..1113611c41a9601f6bd4395d91317b263ad32980 --- /dev/null +++ b/precise_coder/4b_rl_k1.0_f1ur/global_step_328/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86081b183f238cea33f9d87c51d2863dbcee7102870af03856afacd31c0f166c +size 7316 diff --git a/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_0.pt b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..bce986825725742978b1df25c6701046ee0ce43a --- /dev/null +++ b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6fd59f850bdffc774a55af8980d6af545d47cf46d42ffc4db823887d08c6e37b +size 14973 diff --git a/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_1.pt b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..c051a243a87769436d17cf673d679554240ba3e3 --- /dev/null +++ b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06527eca72f06fc8a5ada369424675f17fea42a342475c7895214989555c2290 +size 15037 diff --git a/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_2.pt b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..fdac9c9c99e1bdded0e9ab94b9afffcb645aaa43 --- /dev/null +++ b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fd251590c946f912085c95dcb85c8913378f50bf577c0fb6979f3ab14fd0a8b4 +size 14973 diff --git a/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_3.pt b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..acd77d9d134a682b806657e4b300450d87cc137f --- /dev/null +++ b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/actor/extra_state_world_size_4_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:96669ab1eb30fe1f1ab62f74c543756848f9aa94b1486d841e148164df3350e8 +size 14909 diff --git a/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/data.pt b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..d154c2b19c66a136e0b1a726b09543dcbb399c9f --- /dev/null +++ b/precise_coder/qwen35-4b-GRPO_unit_bs96_bypass/global_step_108/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38f1982c6fbfb4b37ef643d943d296615c88e54f550b750fe0a56235f69e8bb1 +size 7316 diff --git a/precise_coder_sft/4b-SFT-27b-1ep/global_step_45/data_0.pt b/precise_coder_sft/4b-SFT-27b-1ep/global_step_45/data_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..d1356a2ae4d1bab5ef96e0ee030a3ed84f005446 --- /dev/null +++ b/precise_coder_sft/4b-SFT-27b-1ep/global_step_45/data_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad515a2805e2adab011801f4577ed3e9292a514ae23b458f25e39a705718805c +size 1959 diff --git a/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/data_1.pt b/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/data_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..7a455461ad050df4e8bd97f637f13ce4d656b3a9 --- /dev/null +++ b/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/data_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:17d46d2fcfd00fd514d41c44d9e40af9835d20b49194a0851ca85c2bccceac26 +size 1959 diff --git a/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-27b-hard-2ep/global_step_22/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-dsv4pro-1ep/global_step_41/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-dsv4pro-1ep/global_step_41/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-dsv4pro-1ep/global_step_41/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-dsv4pro-hard-2ep/global_step_21/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-dsv4pro-hard-2ep/global_step_21/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-dsv4pro-hard-2ep/global_step_21/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/data_1.pt b/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/data_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..717f909d37991d48dd7b5c59078d135666d5b99a --- /dev/null +++ b/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/data_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d647e6c1be88f0c6b340727fc31eef7e1e80107f34418cd4f7c5397827ff82b4 +size 1959 diff --git a/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-glm-1ep/global_step_31/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/data_1.pt b/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/data_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..248d12f32533b672c6972b1d2635255fcfbac1f2 --- /dev/null +++ b/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/data_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:096936bc1ed7074e3cee4e238930c7ef6ffc5c6c976432a5e859610c1e1c627a +size 1959 diff --git a/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-glm-hard-2ep/global_step_21/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-rej32-think7k/global_step_29/data_1.pt b/precise_coder_sft/4b-SFT-rej32-think7k/global_step_29/data_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..7e42973a9ea7695f20c3c5ff9647dfad44a5a3a3 --- /dev/null +++ b/precise_coder_sft/4b-SFT-rej32-think7k/global_step_29/data_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:68caf357a86eedead8850bec98d3d9efcdcfed443d9ef8c39a139e001b827cc9 +size 1959 diff --git a/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_0.pt b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..bfef1a7d8e743c081e65a6773540352c15498a13 --- /dev/null +++ b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acfd56970badd199fafc2c9731e5535c3341ab013a90b048466cfcb62e679bfd +size 1959 diff --git a/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_1.pt b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..5fd227a975991620152b0361ff1c41982c750733 --- /dev/null +++ b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/data_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e285f4f5f8dcc8e3ed9b6088f2e0a51894a461a19d3030e33ed5471902ade558 +size 1959 diff --git a/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-rej32-trackC/global_step_18/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325 diff --git a/precise_coder_sft/4b-SFT-rej32/global_step_32/huggingface/tokenizer.json b/precise_coder_sft/4b-SFT-rej32/global_step_32/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5520bfd2dd834ce386c1312c410fa71af56db5ad --- /dev/null +++ b/precise_coder_sft/4b-SFT-rej32/global_step_32/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523 +size 19989325