/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(

==================================================
Evaluating model: phyx
Path: /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final
GPUs: [0, 1, 2, 3]
==================================================
Loaded 1533 test samples
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 0] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 1] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 2] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 3] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final...
`torch_dtype` is deprecated! Use `dtype` instead!

Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.16s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.22s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.49s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.52s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.20it/s]
Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.14it/s]
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.

Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.13it/s]
Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.07it/s]
[GPU 2] Processing 383 samples...
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.

Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.06s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.12s/it]

Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.09s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.16s/it]
[GPU 3] Processing 383 samples...
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
[GPU 1] Processing 383 samples...
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
[GPU 0] Processing 384 samples...
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
[GPU 3] 50/383 done
[GPU 1] 50/383 done
[GPU 0] 50/384 done
[GPU 2] 50/383 done
[GPU 1] 100/383 done
[GPU 2] 100/383 done
[GPU 3] 100/383 done
[GPU 0] 100/384 done
[GPU 1] 150/383 done
[GPU 3] 150/383 done
[GPU 2] 150/383 done
[GPU 0] 150/384 done
[GPU 1] 200/383 done
[GPU 3] 200/383 done
[GPU 2] 200/383 done
[GPU 0] 200/384 done
[GPU 3] 250/383 done
[GPU 1] 250/383 done
[GPU 2] 250/383 done
[GPU 0] 250/384 done
[GPU 3] 300/383 done
[GPU 1] 300/383 done
[GPU 2] 300/383 done
[GPU 0] 300/384 done
[GPU 1] 350/383 done
[GPU 3] 350/383 done
[GPU 2] 350/383 done
[GPU 1] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx.jsonl_gpu1.jsonl
[GPU 3] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx.jsonl_gpu3.jsonl
[GPU 0] 350/384 done
[GPU 2] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx.jsonl_gpu2.jsonl
[GPU 0] Saved 384 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx.jsonl_gpu0.jsonl
Merged 1533 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx.jsonl

==================================================
Evaluating model: phyx_50000
Path: /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final
GPUs: [4, 5, 6, 7]
==================================================
Loaded 1533 test samples
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 4] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 5] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 6] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final...
`torch_dtype` is deprecated! Use `dtype` instead!
/opt/conda/lib/python3.11/site-packages/torch/utils/_pytree.py:185: FutureWarning: optree is installed but the version is too old to support PyTorch Dynamo in C++ pytree. C++ pytree support is disabled. Please consider upgrading optree using `python3 -m pip install --upgrade 'optree>=0.13.0'`.
  warnings.warn(
[GPU 7] Loading model from /workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final...
`torch_dtype` is deprecated! Use `dtype` instead!

Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.26s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.39s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.43s/it]
Loading checkpoint shards:  50%|█████     | 1/2 [00:01<00:01,  1.44s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.08it/s]
Loading checkpoint shards: 100%|██████████| 2/2 [00:01<00:00,  1.02it/s]
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.

Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.02s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.08s/it]

Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.09s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.14s/it]

Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.09s/it]
Loading checkpoint shards: 100%|██████████| 2/2 [00:02<00:00,  1.14s/it]
[GPU 7] Processing 383 samples...
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
The tokenizer you are loading from '/workspace/rl4phyx/RL4Phyx/SFT/checkpoints/sft_qwen25vl_3b_fullft_phyx_50000/final' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
[GPU 4] Processing 384 samples...
[GPU 6] Processing 383 samples...
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
[GPU 5] Processing 383 samples...
The following generation flags are not valid and may be ignored: ['temperature']. Set `TRANSFORMERS_VERBOSITY=info` for more details.
[GPU 6] 50/383 done
[GPU 7] 50/383 done
[GPU 4] 50/384 done
[GPU 5] 50/383 done
[GPU 6] 100/383 done
[GPU 5] 100/383 done
[GPU 4] 100/384 done
[GPU 7] 100/383 done
[GPU 5] 150/383 done
[GPU 4] 150/384 done
[GPU 6] 150/383 done
[GPU 7] 150/383 done
[GPU 5] 200/383 done
[GPU 4] 200/384 done
[GPU 6] 200/383 done
[GPU 7] 200/383 done
[GPU 5] 250/383 done
[GPU 6] 250/383 done
[GPU 4] 250/384 done
[GPU 7] 250/383 done
[GPU 5] 300/383 done
[GPU 4] 300/384 done
[GPU 6] 300/383 done
[GPU 7] 300/383 done
[GPU 5] 350/383 done
[GPU 4] 350/384 done
[GPU 7] 350/383 done
[GPU 6] 350/383 done
[GPU 5] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx_50000.jsonl_gpu5.jsonl
[GPU 4] Saved 384 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx_50000.jsonl_gpu4.jsonl
[GPU 6] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx_50000.jsonl_gpu6.jsonl
[GPU 7] Saved 383 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx_50000.jsonl_gpu7.jsonl
Merged 1533 results to /workspace/rl4phyx/RL4Phyx/SFT/sft_eval_footprint/inference_results_phyx_50000.jsonl

ALL_EVAL_DONE