/home/msyu/workspace/venv/bin/python -m mlc_llm gen_config /tmp/tmpy7oe9l51/repo --quantization q0f32 --conv-template phi-3 --output /tmp/tmptqlm3cdw [2024-05-08 13:34:15] INFO auto_config.py:115: Found model configuration: /tmp/tmpy7oe9l51/repo/config.json [2024-05-08 13:34:15] INFO auto_config.py:153: Found model type: phi3. Use `--model-type` to override. [2024-05-08 13:34:15] INFO phi3_model.py:53: context_window_size not found in config.json. Falling back to max_position_embeddings (4096) [2024-05-08 13:34:15] INFO phi3_model.py:68: prefill_chunk_size defaults to 2048 [2024-05-08 13:34:15] INFO config.py:106: Overriding max_batch_size from 1 to 80 [2024-05-08 13:34:15] INFO gen_config.py:255: [generation_config.json] Setting bos_token_id: 1 [2024-05-08 13:34:15] INFO gen_config.py:255: [generation_config.json] Setting eos_token_id: [32000, 32001, 32007] [2024-05-08 13:34:15] INFO gen_config.py:255: [generation_config.json] Setting pad_token_id: 32000 [2024-05-08 13:34:15] INFO gen_config.py:267: Found tokenizer config: /tmp/tmpy7oe9l51/repo/tokenizer.model. Copying to /tmp/tmptqlm3cdw/tokenizer.model [2024-05-08 13:34:15] INFO gen_config.py:267: Found tokenizer config: /tmp/tmpy7oe9l51/repo/tokenizer.json. Copying to /tmp/tmptqlm3cdw/tokenizer.json [2024-05-08 13:34:15] INFO gen_config.py:269: Not found tokenizer config: /tmp/tmpy7oe9l51/repo/vocab.json [2024-05-08 13:34:15] INFO gen_config.py:269: Not found tokenizer config: /tmp/tmpy7oe9l51/repo/merges.txt [2024-05-08 13:34:15] INFO gen_config.py:267: Found tokenizer config: /tmp/tmpy7oe9l51/repo/added_tokens.json. Copying to /tmp/tmptqlm3cdw/added_tokens.json [2024-05-08 13:34:15] INFO gen_config.py:267: Found tokenizer config: /tmp/tmpy7oe9l51/repo/tokenizer_config.json. Copying to /tmp/tmptqlm3cdw/tokenizer_config.json [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting temperature: 0.7 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting presence_penalty: 0.0 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting frequency_penalty: 0.0 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting repetition_penalty: 1.0 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting top_p: 0.95 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting mean_gen_len: 128 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting max_gen_len: 512 [2024-05-08 13:34:15] INFO gen_config.py:80: [System default] Setting shift_fill_factor: 0.3 [2024-05-08 13:34:15] INFO gen_config.py:335: Dumping configuration file to: /tmp/tmptqlm3cdw/mlc-chat-config.json /home/msyu/workspace/venv/bin/python -m mlc_llm convert_weight /tmp/tmpy7oe9l51/repo --quantization q0f32 --source-format auto --output /tmp/tmptqlm3cdw [2024-05-08 13:34:16] INFO auto_config.py:115: Found model configuration: /tmp/tmpy7oe9l51/repo/config.json [2024-05-08 13:34:16] INFO auto_device.py:79: Found device: cuda:0 [2024-05-08 13:34:17] INFO auto_device.py:88: Not found device: rocm:0 [2024-05-08 13:34:18] INFO auto_device.py:88: Not found device: metal:0 [2024-05-08 13:34:18] INFO auto_device.py:88: Not found device: vulkan:0 [2024-05-08 13:34:19] INFO auto_device.py:79: Found device: opencl:0 [2024-05-08 13:34:19] INFO auto_device.py:35: Using device: cuda:0 [2024-05-08 13:34:19] INFO auto_weight.py:70: Finding weights in: /tmp/tmpy7oe9l51/repo [2024-05-08 13:34:19] INFO auto_weight.py:136: Not found Huggingface PyTorch [2024-05-08 13:34:19] INFO auto_weight.py:143: Found source weight format: huggingface-safetensor. Source configuration: /tmp/tmpy7oe9l51/repo/model.safetensors.index.json [2024-05-08 13:34:19] INFO auto_weight.py:106: Using source weight configuration: /tmp/tmpy7oe9l51/repo/model.safetensors.index.json. Use `--source` to override. [2024-05-08 13:34:19] INFO auto_weight.py:110: Using source weight format: huggingface-safetensor. Use `--source-format` to override. [2024-05-08 13:34:19] INFO auto_config.py:153: Found model type: phi3. Use `--model-type` to override. [2024-05-08 13:34:19] INFO phi3_model.py:53: context_window_size not found in config.json. Falling back to max_position_embeddings (4096) [2024-05-08 13:34:19] INFO phi3_model.py:68: prefill_chunk_size defaults to 2048 Weight conversion with arguments: --config /tmp/tmpy7oe9l51/repo/config.json --quantization NoQuantize(name='q0f32', kind='no-quant', model_dtype='float32') --model-type phi3 --device cuda:0 --source /tmp/tmpy7oe9l51/repo/model.safetensors.index.json --source-format huggingface-safetensor --output /tmp/tmptqlm3cdw Start storing to cache /tmp/tmptqlm3cdw 0%| | 0/195 [00:00 type is zero. setattr(self, word, getattr(machar, word).flat[0]) /home/msyu/workspace/venv/lib/python3.10/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for type is zero. return self._float_to_str(self.smallest_subnormal) 1%| | 1/195 [00:02<07:35, 2.35s/it] [2024-05-08 13:34:22] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.ln.weight", shape: (3072,), dtype: float32 1%| | 1/195 [00:02<07:35, 2.35s/it] [2024-05-08 13:34:22] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 1%| | 1/195 [00:02<07:35, 2.35s/it] 2%|▏ | 3/195 [00:02<02:14, 1.43it/s] [2024-05-08 13:34:23] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 2%|▏ | 3/195 [00:02<02:14, 1.43it/s] 2%|▏ | 4/195 [00:03<01:58, 1.61it/s] [2024-05-08 13:34:23] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.post_attention_layernorm.weight", shape: (3072,), dtype: float32 2%|▏ | 4/195 [00:03<01:58, 1.61it/s] [2024-05-08 13:34:23] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 2%|▏ | 4/195 [00:03<01:58, 1.61it/s] 3%|▎ | 6/195 [00:03<01:11, 2.64it/s] [2024-05-08 13:34:23] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.ln.weight", shape: (3072,), dtype: float32 3%|▎ | 6/195 [00:03<01:11, 2.64it/s] [2024-05-08 13:34:23] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 3%|▎ | 6/195 [00:03<01:11, 2.64it/s] 4%|▍ | 8/195 [00:03<00:50, 3.69it/s] [2024-05-08 13:34:24] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 4%|▍ | 8/195 [00:03<00:50, 3.69it/s] 5%|▍ | 9/195 [00:04<00:58, 3.16it/s] [2024-05-08 13:34:24] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.post_attention_layernorm.weight", shape: (3072,), dtype: float32 5%|▍ | 9/195 [00:04<00:58, 3.16it/s] [2024-05-08 13:34:24] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 5%|▍ | 9/195 [00:04<00:58, 3.16it/s] [2024-05-08 13:34:24] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.22.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 5%|▍ | 9/195 [00:04<00:58, 3.16it/s] 6%|▌ | 12/195 [00:04<00:39, 4.64it/s] [2024-05-08 13:34:24] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.ln.weight", shape: (3072,), dtype: float32 6%|▌ | 12/195 [00:04<00:39, 4.64it/s] [2024-05-08 13:34:25] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 6%|▌ | 12/195 [00:04<00:39, 4.64it/s] 7%|▋ | 14/195 [00:04<00:33, 5.41it/s] [2024-05-08 13:34:25] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 7%|▋ | 14/195 [00:04<00:33, 5.41it/s] 8%|▊ | 15/195 [00:05<00:42, 4.21it/s] [2024-05-08 13:34:25] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.post_attention_layernorm.weight", shape: (3072,), dtype: float32 8%|▊ | 15/195 [00:05<00:42, 4.21it/s] [2024-05-08 13:34:25] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 8%|▊ | 15/195 [00:05<00:42, 4.21it/s] [2024-05-08 13:34:25] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.23.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 8%|▊ | 15/195 [00:05<00:42, 4.21it/s] 9%|▉ | 18/195 [00:05<00:32, 5.42it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.ln.weight", shape: (3072,), dtype: float32 9%|▉ | 18/195 [00:05<00:32, 5.42it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 9%|▉ | 18/195 [00:05<00:32, 5.42it/s] 10%|█ | 20/195 [00:05<00:28, 6.06it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 10%|█ | 20/195 [00:05<00:28, 6.06it/s] 11%|█ | 21/195 [00:06<00:37, 4.59it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.post_attention_layernorm.weight", shape: (3072,), dtype: float32 11%|█ | 21/195 [00:06<00:37, 4.59it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 11%|█ | 21/195 [00:06<00:37, 4.59it/s] [2024-05-08 13:34:26] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.24.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 11%|█ | 21/195 [00:06<00:37, 4.59it/s] 12%|█▏ | 24/195 [00:06<00:29, 5.72it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.ln.weight", shape: (3072,), dtype: float32 12%|█▏ | 24/195 [00:06<00:29, 5.72it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 12%|█▏ | 24/195 [00:06<00:29, 5.72it/s] 13%|█▎ | 26/195 [00:06<00:26, 6.30it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 13%|█▎ | 26/195 [00:06<00:26, 6.30it/s] 14%|█▍ | 27/195 [00:07<00:37, 4.54it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.post_attention_layernorm.weight", shape: (3072,), dtype: float32 14%|█▍ | 27/195 [00:07<00:37, 4.54it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 14%|█▍ | 27/195 [00:07<00:37, 4.54it/s] [2024-05-08 13:34:27] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.25.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 14%|█▍ | 27/195 [00:07<00:37, 4.54it/s] 15%|█▌ | 30/195 [00:07<00:29, 5.66it/s] [2024-05-08 13:34:28] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.ln.weight", shape: (3072,), dtype: float32 15%|█▌ | 30/195 [00:07<00:29, 5.66it/s] [2024-05-08 13:34:28] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 15%|█▌ | 30/195 [00:07<00:29, 5.66it/s] 16%|█▋ | 32/195 [00:07<00:26, 6.24it/s] [2024-05-08 13:34:28] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 16%|█▋ | 32/195 [00:07<00:26, 6.24it/s] 17%|█▋ | 33/195 [00:08<00:36, 4.46it/s] [2024-05-08 13:34:28] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.post_attention_layernorm.weight", shape: (3072,), dtype: float32 17%|█▋ | 33/195 [00:08<00:36, 4.46it/s] [2024-05-08 13:34:29] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 17%|█▋ | 33/195 [00:08<00:36, 4.46it/s] [2024-05-08 13:34:29] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.26.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 17%|█▋ | 33/195 [00:08<00:36, 4.46it/s] 18%|█▊ | 36/195 [00:08<00:28, 5.59it/s] [2024-05-08 13:34:29] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.ln.weight", shape: (3072,), dtype: float32 18%|█▊ | 36/195 [00:08<00:28, 5.59it/s] [2024-05-08 13:34:29] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 18%|█▊ | 36/195 [00:08<00:28, 5.59it/s] 19%|█▉ | 38/195 [00:09<00:30, 5.13it/s] [2024-05-08 13:34:29] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 19%|█▉ | 38/195 [00:09<00:30, 5.13it/s] 20%|██ | 39/195 [00:09<00:39, 3.90it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.post_attention_layernorm.weight", shape: (3072,), dtype: float32 20%|██ | 39/195 [00:09<00:39, 3.90it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 20%|██ | 39/195 [00:09<00:39, 3.90it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.27.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 20%|██ | 39/195 [00:09<00:39, 3.90it/s] 22%|██▏ | 42/195 [00:10<00:30, 5.06it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.ln.weight", shape: (3072,), dtype: float32 22%|██▏ | 42/195 [00:10<00:30, 5.06it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 22%|██▏ | 42/195 [00:10<00:30, 5.06it/s] 23%|██▎ | 44/195 [00:10<00:26, 5.71it/s] [2024-05-08 13:34:30] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 23%|██▎ | 44/195 [00:10<00:26, 5.71it/s] 23%|██▎ | 45/195 [00:10<00:35, 4.21it/s] [2024-05-08 13:34:31] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.post_attention_layernorm.weight", shape: (3072,), dtype: float32 23%|██▎ | 45/195 [00:10<00:35, 4.21it/s] [2024-05-08 13:34:31] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 23%|██▎ | 45/195 [00:10<00:35, 4.21it/s] [2024-05-08 13:34:31] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.28.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 23%|██▎ | 45/195 [00:10<00:35, 4.21it/s] 25%|██▍ | 48/195 [00:11<00:27, 5.34it/s] [2024-05-08 13:34:31] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.ln.weight", shape: (3072,), dtype: float32 25%|██▍ | 48/195 [00:11<00:27, 5.34it/s] [2024-05-08 13:34:31] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 25%|██▍ | 48/195 [00:11<00:27, 5.34it/s] 26%|██▌ | 50/195 [00:11<00:24, 5.96it/s] [2024-05-08 13:34:32] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 26%|██▌ | 50/195 [00:11<00:24, 5.96it/s] 26%|██▌ | 51/195 [00:12<00:33, 4.29it/s] [2024-05-08 13:34:32] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.post_attention_layernorm.weight", shape: (3072,), dtype: float32 26%|██▌ | 51/195 [00:12<00:33, 4.29it/s] [2024-05-08 13:34:32] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 26%|██▌ | 51/195 [00:12<00:33, 4.29it/s] [2024-05-08 13:34:32] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.29.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 26%|██▌ | 51/195 [00:12<00:33, 4.29it/s] 28%|██▊ | 54/195 [00:12<00:25, 5.43it/s] [2024-05-08 13:34:32] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.ln.weight", shape: (3072,), dtype: float32 28%|██▊ | 54/195 [00:12<00:25, 5.43it/s] [2024-05-08 13:34:33] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 28%|██▊ | 54/195 [00:12<00:25, 5.43it/s] 29%|██▊ | 56/195 [00:12<00:23, 6.03it/s] [2024-05-08 13:34:33] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 29%|██▊ | 56/195 [00:12<00:23, 6.03it/s] 29%|██▉ | 57/195 [00:13<00:32, 4.28it/s] [2024-05-08 13:34:33] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.post_attention_layernorm.weight", shape: (3072,), dtype: float32 29%|██▉ | 57/195 [00:13<00:32, 4.28it/s] [2024-05-08 13:34:33] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 29%|██▉ | 57/195 [00:13<00:32, 4.28it/s] [2024-05-08 13:34:33] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.30.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 29%|██▉ | 57/195 [00:13<00:32, 4.28it/s] 31%|███ | 60/195 [00:13<00:24, 5.42it/s] [2024-05-08 13:34:34] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.ln.weight", shape: (3072,), dtype: float32 31%|███ | 60/195 [00:13<00:24, 5.42it/s] [2024-05-08 13:34:34] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 31%|███ | 60/195 [00:13<00:24, 5.42it/s] 32%|███▏ | 62/195 [00:13<00:22, 6.02it/s] [2024-05-08 13:34:34] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 32%|███▏ | 62/195 [00:13<00:22, 6.02it/s] 32%|███▏ | 63/195 [00:14<00:44, 2.96it/s] [2024-05-08 13:34:35] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.post_attention_layernorm.weight", shape: (3072,), dtype: float32 32%|███▏ | 63/195 [00:14<00:44, 2.96it/s] [2024-05-08 13:34:35] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 32%|███▏ | 63/195 [00:14<00:44, 2.96it/s] [2024-05-08 13:34:35] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.31.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 32%|███▏ | 63/195 [00:15<00:44, 2.96it/s] 34%|███▍ | 66/195 [00:15<00:31, 4.05it/s] [2024-05-08 13:34:35] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.norm.weight", shape: (3072,), dtype: float32 34%|███▍ | 66/195 [00:15<00:31, 4.05it/s] [2024-05-08 13:34:35] INFO huggingface_loader.py:196: Unloading HF weight file: /tmp/tmpy7oe9l51/repo/model-00002-of-00002.safetensors 34%|███▍ | 66/195 [00:15<00:31, 4.05it/s] [2024-05-08 13:34:36] INFO huggingface_loader.py:184: Loading HF parameters from: /tmp/tmpy7oe9l51/repo/model-00001-of-00002.safetensors 34%|███▍ | 66/195 [00:15<00:31, 4.05it/s] [2024-05-08 13:34:36] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.embd.weight", shape: (32064, 3072), dtype: float32 34%|███▍ | 66/195 [00:16<00:31, 4.05it/s] 35%|███▍ | 68/195 [00:17<01:01, 2.08it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.ln.weight", shape: (3072,), dtype: float32 35%|███▍ | 68/195 [00:17<01:01, 2.08it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 35%|███▍ | 68/195 [00:17<01:01, 2.08it/s] 36%|███▌ | 70/195 [00:17<00:47, 2.65it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 36%|███▌ | 70/195 [00:17<00:47, 2.65it/s] 36%|███▋ | 71/195 [00:18<00:48, 2.55it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.post_attention_layernorm.weight", shape: (3072,), dtype: float32 36%|███▋ | 71/195 [00:18<00:48, 2.55it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 36%|███▋ | 71/195 [00:18<00:48, 2.55it/s] [2024-05-08 13:34:38] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.0.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 36%|███▋ | 71/195 [00:18<00:48, 2.55it/s] 38%|███▊ | 74/195 [00:18<00:33, 3.61it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.ln.weight", shape: (3072,), dtype: float32 38%|███▊ | 74/195 [00:18<00:33, 3.61it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 38%|███▊ | 74/195 [00:18<00:33, 3.61it/s] 39%|███▉ | 76/195 [00:18<00:27, 4.32it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 39%|███▉ | 76/195 [00:18<00:27, 4.32it/s] 39%|███▉ | 77/195 [00:19<00:32, 3.67it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.post_attention_layernorm.weight", shape: (3072,), dtype: float32 39%|███▉ | 77/195 [00:19<00:32, 3.67it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 39%|███▉ | 77/195 [00:19<00:32, 3.67it/s] [2024-05-08 13:34:39] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.1.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 39%|███▉ | 77/195 [00:19<00:32, 3.67it/s] 41%|████ | 80/195 [00:19<00:23, 4.82it/s] [2024-05-08 13:34:40] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.ln.weight", shape: (3072,), dtype: float32 41%|████ | 80/195 [00:19<00:23, 4.82it/s] [2024-05-08 13:34:40] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 41%|████ | 80/195 [00:19<00:23, 4.82it/s] 42%|████▏ | 82/195 [00:19<00:20, 5.49it/s] [2024-05-08 13:34:40] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 42%|████▏ | 82/195 [00:19<00:20, 5.49it/s] 43%|████▎ | 83/195 [00:20<00:25, 4.33it/s] [2024-05-08 13:34:40] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.post_attention_layernorm.weight", shape: (3072,), dtype: float32 43%|████▎ | 83/195 [00:20<00:25, 4.33it/s] [2024-05-08 13:34:40] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 43%|████▎ | 83/195 [00:20<00:25, 4.33it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.10.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 43%|████▎ | 83/195 [00:20<00:25, 4.33it/s] 44%|████▍ | 86/195 [00:20<00:20, 5.45it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.ln.weight", shape: (3072,), dtype: float32 44%|████▍ | 86/195 [00:20<00:20, 5.45it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 44%|████▍ | 86/195 [00:20<00:20, 5.45it/s] 45%|████▌ | 88/195 [00:20<00:17, 6.06it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 45%|████▌ | 88/195 [00:20<00:17, 6.06it/s] 46%|████▌ | 89/195 [00:21<00:22, 4.62it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.post_attention_layernorm.weight", shape: (3072,), dtype: float32 46%|████▌ | 89/195 [00:21<00:22, 4.62it/s] [2024-05-08 13:34:41] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 46%|████▌ | 89/195 [00:21<00:22, 4.62it/s] [2024-05-08 13:34:42] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.11.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 46%|████▌ | 89/195 [00:21<00:22, 4.62it/s] 47%|████▋ | 92/195 [00:21<00:18, 5.71it/s] [2024-05-08 13:34:42] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.ln.weight", shape: (3072,), dtype: float32 47%|████▋ | 92/195 [00:21<00:18, 5.71it/s] [2024-05-08 13:34:42] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 47%|████▋ | 92/195 [00:21<00:18, 5.71it/s] 48%|████▊ | 94/195 [00:21<00:16, 6.29it/s] [2024-05-08 13:34:42] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 48%|████▊ | 94/195 [00:21<00:16, 6.29it/s] 49%|████▊ | 95/195 [00:22<00:21, 4.69it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.post_attention_layernorm.weight", shape: (3072,), dtype: float32 49%|████▊ | 95/195 [00:22<00:21, 4.69it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 49%|████▊ | 95/195 [00:22<00:21, 4.69it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.12.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 49%|████▊ | 95/195 [00:22<00:21, 4.69it/s] 50%|█████ | 98/195 [00:22<00:16, 5.77it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.ln.weight", shape: (3072,), dtype: float32 50%|█████ | 98/195 [00:22<00:16, 5.77it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 50%|█████ | 98/195 [00:22<00:16, 5.77it/s] 51%|█████▏ | 100/195 [00:23<00:14, 6.34it/s] [2024-05-08 13:34:43] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 51%|█████▏ | 100/195 [00:23<00:14, 6.34it/s] 52%|█████▏ | 101/195 [00:23<00:20, 4.67it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.post_attention_layernorm.weight", shape: (3072,), dtype: float32 52%|█████▏ | 101/195 [00:23<00:20, 4.67it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 52%|█████▏ | 101/195 [00:23<00:20, 4.67it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.13.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 52%|█████▏ | 101/195 [00:23<00:20, 4.67it/s] 53%|█████▎ | 104/195 [00:23<00:15, 5.75it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.ln.weight", shape: (3072,), dtype: float32 53%|█████▎ | 104/195 [00:23<00:15, 5.75it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 53%|█████▎ | 104/195 [00:23<00:15, 5.75it/s] 54%|█████▍ | 106/195 [00:24<00:14, 6.33it/s] [2024-05-08 13:34:44] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 54%|█████▍ | 106/195 [00:24<00:14, 6.33it/s] 55%|█████▍ | 107/195 [00:24<00:18, 4.69it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.post_attention_layernorm.weight", shape: (3072,), dtype: float32 55%|█████▍ | 107/195 [00:24<00:18, 4.69it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 55%|█████▍ | 107/195 [00:24<00:18, 4.69it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.14.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 55%|█████▍ | 107/195 [00:24<00:18, 4.69it/s] 56%|█████▋ | 110/195 [00:24<00:14, 5.77it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.ln.weight", shape: (3072,), dtype: float32 56%|█████▋ | 110/195 [00:24<00:14, 5.77it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 56%|█████▋ | 110/195 [00:24<00:14, 5.77it/s] 57%|█████▋ | 112/195 [00:25<00:13, 6.34it/s] [2024-05-08 13:34:45] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 57%|█████▋ | 112/195 [00:25<00:13, 6.34it/s] 58%|█████▊ | 113/195 [00:25<00:17, 4.65it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.post_attention_layernorm.weight", shape: (3072,), dtype: float32 58%|█████▊ | 113/195 [00:25<00:17, 4.65it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 58%|█████▊ | 113/195 [00:25<00:17, 4.65it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.15.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 58%|█████▊ | 113/195 [00:25<00:17, 4.65it/s] 59%|█████▉ | 116/195 [00:26<00:13, 5.70it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.ln.weight", shape: (3072,), dtype: float32 59%|█████▉ | 116/195 [00:26<00:13, 5.70it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 59%|█████▉ | 116/195 [00:26<00:13, 5.70it/s] 61%|██████ | 118/195 [00:26<00:12, 6.22it/s] [2024-05-08 13:34:46] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 61%|██████ | 118/195 [00:26<00:12, 6.22it/s] 61%|██████ | 119/195 [00:26<00:17, 4.47it/s] [2024-05-08 13:34:47] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.post_attention_layernorm.weight", shape: (3072,), dtype: float32 61%|██████ | 119/195 [00:26<00:17, 4.47it/s] [2024-05-08 13:34:47] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 61%|██████ | 119/195 [00:26<00:17, 4.47it/s] [2024-05-08 13:34:47] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.16.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 61%|██████ | 119/195 [00:26<00:17, 4.47it/s] 63%|██████▎ | 122/195 [00:27<00:13, 5.57it/s] [2024-05-08 13:34:47] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.ln.weight", shape: (3072,), dtype: float32 63%|██████▎ | 122/195 [00:27<00:13, 5.57it/s] [2024-05-08 13:34:47] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 63%|██████▎ | 122/195 [00:27<00:13, 5.57it/s] 64%|██████▎ | 124/195 [00:27<00:11, 6.17it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 64%|██████▎ | 124/195 [00:27<00:11, 6.17it/s] 64%|██████▍ | 125/195 [00:27<00:15, 4.56it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.post_attention_layernorm.weight", shape: (3072,), dtype: float32 64%|██████▍ | 125/195 [00:27<00:15, 4.56it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 64%|██████▍ | 125/195 [00:27<00:15, 4.56it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.17.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 64%|██████▍ | 125/195 [00:28<00:15, 4.56it/s] 66%|██████▌ | 128/195 [00:28<00:11, 5.65it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.ln.weight", shape: (3072,), dtype: float32 66%|██████▌ | 128/195 [00:28<00:11, 5.65it/s] [2024-05-08 13:34:48] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 66%|██████▌ | 128/195 [00:28<00:11, 5.65it/s] 67%|██████▋ | 130/195 [00:28<00:10, 6.24it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 67%|██████▋ | 130/195 [00:28<00:10, 6.24it/s] 67%|██████▋ | 131/195 [00:29<00:13, 4.60it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.post_attention_layernorm.weight", shape: (3072,), dtype: float32 67%|██████▋ | 131/195 [00:29<00:13, 4.60it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 67%|██████▋ | 131/195 [00:29<00:13, 4.60it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.18.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 67%|██████▋ | 131/195 [00:29<00:13, 4.60it/s] 69%|██████▊ | 134/195 [00:29<00:10, 5.69it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.ln.weight", shape: (3072,), dtype: float32 69%|██████▊ | 134/195 [00:29<00:10, 5.69it/s] [2024-05-08 13:34:49] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 69%|██████▊ | 134/195 [00:29<00:10, 5.69it/s] 70%|██████▉ | 136/195 [00:29<00:09, 6.28it/s] [2024-05-08 13:34:50] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 70%|██████▉ | 136/195 [00:29<00:09, 6.28it/s] 70%|███████ | 137/195 [00:30<00:13, 4.27it/s] [2024-05-08 13:34:50] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.post_attention_layernorm.weight", shape: (3072,), dtype: float32 70%|███████ | 137/195 [00:30<00:13, 4.27it/s] [2024-05-08 13:34:50] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 70%|███████ | 137/195 [00:30<00:13, 4.27it/s] [2024-05-08 13:34:50] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.19.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 70%|███████ | 137/195 [00:30<00:13, 4.27it/s] 72%|███████▏ | 140/195 [00:30<00:10, 5.33it/s] [2024-05-08 13:34:51] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.ln.weight", shape: (3072,), dtype: float32 72%|███████▏ | 140/195 [00:30<00:10, 5.33it/s] [2024-05-08 13:34:51] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 72%|███████▏ | 140/195 [00:30<00:10, 5.33it/s] 73%|███████▎ | 142/195 [00:30<00:08, 5.93it/s] [2024-05-08 13:34:51] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 73%|███████▎ | 142/195 [00:30<00:08, 5.93it/s] 73%|███████▎ | 143/195 [00:31<00:11, 4.50it/s] [2024-05-08 13:34:51] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.post_attention_layernorm.weight", shape: (3072,), dtype: float32 73%|███████▎ | 143/195 [00:31<00:11, 4.50it/s] [2024-05-08 13:34:51] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 73%|███████▎ | 143/195 [00:31<00:11, 4.50it/s] [2024-05-08 13:34:52] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.2.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 73%|███████▎ | 143/195 [00:31<00:11, 4.50it/s] 75%|███████▍ | 146/195 [00:31<00:08, 5.60it/s] [2024-05-08 13:34:52] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.ln.weight", shape: (3072,), dtype: float32 75%|███████▍ | 146/195 [00:31<00:08, 5.60it/s] [2024-05-08 13:34:52] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 75%|███████▍ | 146/195 [00:31<00:08, 5.60it/s] 76%|███████▌ | 148/195 [00:31<00:07, 6.20it/s] [2024-05-08 13:34:52] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 76%|███████▌ | 148/195 [00:31<00:07, 6.20it/s] 76%|███████▋ | 149/195 [00:32<00:10, 4.39it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.post_attention_layernorm.weight", shape: (3072,), dtype: float32 76%|███████▋ | 149/195 [00:32<00:10, 4.39it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 76%|███████▋ | 149/195 [00:32<00:10, 4.39it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.20.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 76%|███████▋ | 149/195 [00:32<00:10, 4.39it/s] 78%|███████▊ | 152/195 [00:32<00:07, 5.50it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.21.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 78%|███████▊ | 152/195 [00:32<00:07, 5.50it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.ln.weight", shape: (3072,), dtype: float32 78%|███████▊ | 152/195 [00:32<00:07, 5.50it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 78%|███████▊ | 152/195 [00:32<00:07, 5.50it/s] 79%|███████▉ | 155/195 [00:33<00:06, 6.52it/s] [2024-05-08 13:34:53] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 79%|███████▉ | 155/195 [00:33<00:06, 6.52it/s] 80%|████████ | 156/195 [00:33<00:08, 4.86it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.post_attention_layernorm.weight", shape: (3072,), dtype: float32 80%|████████ | 156/195 [00:33<00:08, 4.86it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 80%|████████ | 156/195 [00:33<00:08, 4.86it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.3.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 80%|████████ | 156/195 [00:33<00:08, 4.86it/s] 82%|████████▏ | 159/195 [00:33<00:06, 5.83it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.ln.weight", shape: (3072,), dtype: float32 82%|████████▏ | 159/195 [00:33<00:06, 5.83it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 82%|████████▏ | 159/195 [00:33<00:06, 5.83it/s] 83%|████████▎ | 161/195 [00:34<00:05, 6.30it/s] [2024-05-08 13:34:54] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 83%|████████▎ | 161/195 [00:34<00:05, 6.30it/s] 83%|████████▎ | 162/195 [00:34<00:08, 4.00it/s] [2024-05-08 13:34:55] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.post_attention_layernorm.weight", shape: (3072,), dtype: float32 83%|████████▎ | 162/195 [00:34<00:08, 4.00it/s] [2024-05-08 13:34:55] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 83%|████████▎ | 162/195 [00:34<00:08, 4.00it/s] [2024-05-08 13:34:55] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.4.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 83%|████████▎ | 162/195 [00:35<00:08, 4.00it/s] 85%|████████▍ | 165/195 [00:35<00:05, 5.09it/s] [2024-05-08 13:34:55] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.ln.weight", shape: (3072,), dtype: float32 85%|████████▍ | 165/195 [00:35<00:05, 5.09it/s] [2024-05-08 13:34:55] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 85%|████████▍ | 165/195 [00:35<00:05, 5.09it/s] 86%|████████▌ | 167/195 [00:35<00:04, 5.72it/s] [2024-05-08 13:34:56] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 86%|████████▌ | 167/195 [00:35<00:04, 5.72it/s] 86%|████████▌ | 168/195 [00:36<00:06, 4.25it/s] [2024-05-08 13:34:56] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.post_attention_layernorm.weight", shape: (3072,), dtype: float32 86%|████████▌ | 168/195 [00:36<00:06, 4.25it/s] [2024-05-08 13:34:56] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 86%|████████▌ | 168/195 [00:36<00:06, 4.25it/s] [2024-05-08 13:34:56] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.5.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 86%|████████▌ | 168/195 [00:36<00:06, 4.25it/s] 88%|████████▊ | 171/195 [00:36<00:04, 5.34it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.ln.weight", shape: (3072,), dtype: float32 88%|████████▊ | 171/195 [00:36<00:04, 5.34it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 88%|████████▊ | 171/195 [00:36<00:04, 5.34it/s] 89%|████████▊ | 173/195 [00:36<00:03, 5.95it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 89%|████████▊ | 173/195 [00:36<00:03, 5.95it/s] 89%|████████▉ | 174/195 [00:37<00:04, 4.39it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.post_attention_layernorm.weight", shape: (3072,), dtype: float32 89%|████████▉ | 174/195 [00:37<00:04, 4.39it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 89%|████████▉ | 174/195 [00:37<00:04, 4.39it/s] [2024-05-08 13:34:57] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.6.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 89%|████████▉ | 174/195 [00:37<00:04, 4.39it/s] 91%|█████████ | 177/195 [00:37<00:03, 5.50it/s] [2024-05-08 13:34:58] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.ln.weight", shape: (3072,), dtype: float32 91%|█████████ | 177/195 [00:37<00:03, 5.50it/s] [2024-05-08 13:34:58] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 91%|█████████ | 177/195 [00:37<00:03, 5.50it/s] 92%|█████████▏| 179/195 [00:37<00:02, 6.11it/s] [2024-05-08 13:34:58] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 92%|█████████▏| 179/195 [00:37<00:02, 6.11it/s] 92%|█████████▏| 180/195 [00:38<00:03, 4.52it/s] [2024-05-08 13:34:58] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.post_attention_layernorm.weight", shape: (3072,), dtype: float32 92%|█████████▏| 180/195 [00:38<00:03, 4.52it/s] [2024-05-08 13:34:58] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 92%|█████████▏| 180/195 [00:38<00:03, 4.52it/s] [2024-05-08 13:34:59] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.7.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 92%|█████████▏| 180/195 [00:38<00:03, 4.52it/s] 94%|█████████▍| 183/195 [00:38<00:02, 5.62it/s] [2024-05-08 13:34:59] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.ln.weight", shape: (3072,), dtype: float32 94%|█████████▍| 183/195 [00:38<00:02, 5.62it/s] [2024-05-08 13:34:59] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 94%|█████████▍| 183/195 [00:38<00:02, 5.62it/s] 95%|█████████▍| 185/195 [00:38<00:01, 6.21it/s] [2024-05-08 13:34:59] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 95%|█████████▍| 185/195 [00:38<00:01, 6.21it/s] 95%|█████████▌| 186/195 [00:39<00:02, 4.07it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.post_attention_layernorm.weight", shape: (3072,), dtype: float32 95%|█████████▌| 186/195 [00:39<00:02, 4.07it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 95%|█████████▌| 186/195 [00:39<00:02, 4.07it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.8.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 95%|█████████▌| 186/195 [00:39<00:02, 4.07it/s] 97%|█████████▋| 189/195 [00:39<00:01, 5.21it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.ln.weight", shape: (3072,), dtype: float32 97%|█████████▋| 189/195 [00:39<00:01, 5.21it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.mlp.down_proj.weight", shape: (3072, 8192), dtype: float32 97%|█████████▋| 189/195 [00:39<00:01, 5.21it/s] 98%|█████████▊| 191/195 [00:40<00:00, 5.85it/s] [2024-05-08 13:35:00] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.mlp.gate_up_proj.weight", shape: (16384, 3072), dtype: float32 98%|█████████▊| 191/195 [00:40<00:00, 5.85it/s] 98%|█████████▊| 192/195 [00:40<00:00, 4.37it/s] [2024-05-08 13:35:01] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.post_attention_layernorm.weight", shape: (3072,), dtype: float32 98%|█████████▊| 192/195 [00:40<00:00, 4.37it/s] [2024-05-08 13:35:01] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.mixer.out_proj.weight", shape: (3072, 3072), dtype: float32 98%|█████████▊| 192/195 [00:40<00:00, 4.37it/s] [2024-05-08 13:35:01] INFO huggingface_loader.py:174: [Not quantized] Parameter: "transformer.h.9.mixer.qkv_proj.weight", shape: (9216, 3072), dtype: float32 98%|█████████▊| 192/195 [00:40<00:00, 4.37it/s] 100%|██████████| 195/195 [00:40<00:00, 5.49it/s] 100%|██████████| 195/195 [00:41<00:00, 4.76it/s] [2024-05-08 13:35:01] INFO huggingface_loader.py:196: Unloading HF weight file: /tmp/tmpy7oe9l51/repo/model-00001-of-00002.safetensors [2024-05-08 13:35:01] INFO stats.py:76: Time usage: HF loading: 1.642 sec; Pre-quantization mapping: 2.009 sec; Quantization: 0.000 sec [2024-05-08 13:35:01] INFO stats.py:90: RAM usage: Peak RAM: 9.262 GB. Total bytes loaded from disk: 14.235 GB [2024-05-08 13:35:01] INFO convert_weight.py:155: Parameter size after quantization: 14.235 GB [2024-05-08 13:35:01] INFO convert_weight.py:160: Total parameters: 3,821,079,552 [2024-05-08 13:35:01] INFO convert_weight.py:161: Bits per parameter: 32.000 [2024-05-08 13:35:01] INFO convert_weight.py:166: Saved to directory: /tmp/tmptqlm3cdw All finished, 130 total shards committed, record saved to /tmp/tmptqlm3cdw/ndarray-cache.json Also saved a bf16 record to /tmp/tmptqlm3cdw/ndarray-cache-b16.json