-
Updated Readme, no need of PR since base nightly works on Blackwell pretty well
Verified myself on RTX 6000 Pro
```shell
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339]
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339] █ █ █▄ ▄█
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339] ▄▄ ▄█ █ █ █ ▀▄▀ █ version 0.23.1rc1.dev471+ge312c5cb2
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339] █▄█▀ █ █ █ █ model /app/efs/models/minimax-m3-nvfp4
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339] ▀▀ ▀▀▀▀▀ ▀▀▀▀▀ ▀ ▀
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:339]
(APIServer pid=1) INFO 06-26 06:55:48 [api_utils.py:273] non-default args: {'model_tag': '/app/efs/models/minimax-m3-nvfp4', 'enable_auto_tool_choice': True, 'tool_call_parser': 'minimax_m3', 'model': '/app/efs/models/minimax-m3-nvfp4', 'trust_remote_code': True, 'max_model_len': -1, 'served_model_name': ['minimax-m3'], 'reasoning_parser': 'minimax_m3', 'tensor_parallel_size': 4, 'block_size': 128, 'gpu_memory_utilization'…
-
Add benchmarking parameters (T=1.0, top_p=0.95, max num tokens 65536) and MXFP8 baseline note to Evaluation section
-
refactor: rename shards to model-NNNNN-of-00088 convention; update index weight_map
-
Remove M3 Coding and Agentic Benchmark Results (BF16 baseline) table
-
Copy vocab.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy video_processor.py from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy tokenizer_config.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy tokenizer.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy special_tokens_map.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy processing_minimax.py from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy preprocessor_config.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy model.safetensors.index.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy merges.txt from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy image_processor.py from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy hf_quant_config.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy generation_config.json from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-059.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-058.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-057.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-056.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-055.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-054.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-053.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-052.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-051.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-050.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-049.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-048.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-047.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-046.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-045.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-044.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-043.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-042.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
docs: note vLLM nightly + PR #46380 required for MiniMax-M3 NVFP4 serving
-
Copy experts-layer-041.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-040.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-039.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-038.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-037.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-036.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-035.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-034.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-033.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-032.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-031.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-030.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-029.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-028.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615
-
Copy experts-layer-027.safetensors from nvidia/MiniMax-M3-NVFP4-Preview0615