diff --git "a/logs.txt" "b/logs.txt" new file mode 100644--- /dev/null +++ "b/logs.txt" @@ -0,0 +1,400 @@ +/home/junrushao/micromamba/envs/python311/bin/python -m mlc_chat gen_config /home/junrushao/tmp/tmpmt7e057e/repo --quantization q3f16_1 --conv-template open_hermes_mistral --output /home/junrushao/tmp/tmpfj48bgud --context-window-size 32768 +[2023-12-29 05:52:25] INFO auto_config.py:115: Found model configuration: /home/junrushao/tmp/tmpmt7e057e/repo/config.json +[2023-12-29 05:52:25] INFO auto_config.py:151: Found model type: mistral. Use `--model-type` to override. +[2023-12-29 05:52:25] INFO mistral_model.py:45: context_window_size not found in config.json. Falling back to max_position_embeddings (32768) +[2023-12-29 05:52:25] INFO mistral_model.py:76: Using sliding window attention, setting context_window_size to -1 +[2023-12-29 05:52:25] INFO flags_model_config_override.py:63: Default prefill_chunk_size to context_window_size (32768) because it is not provided +[2023-12-29 05:52:25] INFO flags_model_config_override.py:112: Overriding context_window_size from -1 to 32768 +[2023-12-29 05:52:25] INFO flags_model_config_override.py:112: Overriding prefill_chunk_size from 4096 to 32768 +[2023-12-29 05:52:25] INFO gen_config.py:117: Not found generation_config.json: /home/junrushao/tmp/tmpmt7e057e/repo/generation_config.json +[2023-12-29 05:52:25] INFO gen_config.py:115: [config.json] Setting bos_token_id: 1 +[2023-12-29 05:52:25] INFO gen_config.py:115: [config.json] Setting eos_token_id: 32000 +[2023-12-29 05:52:25] INFO gen_config.py:127: Found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/tokenizer.model. Copying to /home/junrushao/tmp/tmpfj48bgud/tokenizer.model +[2023-12-29 05:52:25] INFO gen_config.py:129: Not found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/tokenizer.json +[2023-12-29 05:52:25] INFO gen_config.py:129: Not found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/vocab.json +[2023-12-29 05:52:25] INFO gen_config.py:129: Not found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/merges.txt +[2023-12-29 05:52:25] INFO gen_config.py:127: Found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/added_tokens.json. Copying to /home/junrushao/tmp/tmpfj48bgud/added_tokens.json +[2023-12-29 05:52:25] INFO gen_config.py:127: Found tokenizer config: /home/junrushao/tmp/tmpmt7e057e/repo/tokenizer_config.json. Copying to /home/junrushao/tmp/tmpfj48bgud/tokenizer_config.json +[2023-12-29 05:52:25] INFO gen_config.py:135: The model has `tokenizer.model` but not `tokenizer.json`. It is always recommended to prefer JSON instead. Attempting to convert using HuggingFace transformers library +Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. +Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. +[2023-12-29 05:52:26] INFO gen_config.py:149: Succesfully converted `tokenizer.model` to: /home/junrushao/tmp/tmpfj48bgud/tokenizer.json +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting pad_token_id: 0 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting temperature: 0.7 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting repetition_penalty: 1.0 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting top_p: 0.95 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting mean_gen_len: 128 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting max_gen_len: 512 +[2023-12-29 05:52:26] INFO gen_config.py:69: [System default] Setting shift_fill_factor: 0.3 +[2023-12-29 05:52:26] INFO gen_config.py:157: Dumping configuration file to: /home/junrushao/tmp/tmpfj48bgud/mlc-chat-config.json +/home/junrushao/micromamba/envs/python311/bin/python -m mlc_chat convert_weight /home/junrushao/tmp/tmpmt7e057e/repo --quantization q3f16_1 --source-format auto --output /home/junrushao/tmp/tmpfj48bgud +[2023-12-29 05:52:27] INFO auto_config.py:115: Found model configuration: /home/junrushao/tmp/tmpmt7e057e/repo/config.json +[2023-12-29 05:52:27] INFO auto_device.py:76: Found device: cuda:0 +[2023-12-29 05:52:27] INFO auto_device.py:76: Found device: cuda:1 +[2023-12-29 05:52:27] INFO auto_device.py:76: Found device: cuda:2 +[2023-12-29 05:52:27] INFO auto_device.py:76: Found device: cuda:3 +[2023-12-29 05:52:27] INFO auto_device.py:85: Not found device: rocm:0 +[2023-12-29 05:52:28] INFO auto_device.py:85: Not found device: metal:0 +[2023-12-29 05:52:28] INFO auto_device.py:85: Not found device: vulkan:0 +[2023-12-29 05:52:29] INFO auto_device.py:85: Not found device: opencl:0 +[2023-12-29 05:52:29] INFO auto_device.py:33: Using device: cuda:0 +[2023-12-29 05:52:29] INFO auto_weight.py:70: Finding weights in: /home/junrushao/tmp/tmpmt7e057e/repo +[2023-12-29 05:52:29] INFO auto_weight.py:120: Found source weight format: huggingface-torch. Source configuration: /home/junrushao/tmp/tmpmt7e057e/repo/pytorch_model.bin.index.json +[2023-12-29 05:52:29] INFO auto_weight.py:149: Not found Huggingface Safetensor +[2023-12-29 05:52:29] INFO auto_weight.py:106: Using source weight configuration: /home/junrushao/tmp/tmpmt7e057e/repo/pytorch_model.bin.index.json. Use `--source` to override. +[2023-12-29 05:52:29] INFO auto_weight.py:110: Using source weight format: huggingface-torch. Use `--source-format` to override. +[2023-12-29 05:52:29] INFO auto_config.py:151: Found model type: mistral. Use `--model-type` to override. +[2023-12-29 05:52:29] INFO mistral_model.py:45: context_window_size not found in config.json. Falling back to max_position_embeddings (32768) +[2023-12-29 05:52:29] INFO mistral_model.py:76: Using sliding window attention, setting context_window_size to -1 +Weight conversion with arguments: + --config /home/junrushao/tmp/tmpmt7e057e/repo/config.json + --quantization GroupQuantize(name='q3f16_1', kind='group-quant', group_size=40, quantize_dtype='int3', storage_dtype='uint32', model_dtype='float16', num_elem_per_storage=10, num_storage_per_group=4, max_int_value=3) + --model-type mistral + --device cuda:0 + --source /home/junrushao/tmp/tmpmt7e057e/repo/pytorch_model.bin.index.json + --source-format huggingface-torch + --output /home/junrushao/tmp/tmpfj48bgud + 0%| | 0/195 [00:00