{ "hf_id": "pfnet/plamo-3-nict-31b-base", "meta": { "title": "PLaMo 3 NICT 31B Base", "slug": "plamo-3-nict-31b-base", "provider": "Preferred Networks", "description": "Largest PLaMo 3 NICT Japanese/English base model with interleaved sliding-window and full attention.", "date_added": "2026-06-03", "date_updated": "2026-09-13", "tasks": [ "text" ], "performance_headline": "31B-class PLaMo 3 base checkpoint for high-quality bilingual generation", "related_recipes": [] }, "recommended_command": { "hardware": "h200", "strategy": "single_node_tp", "variant": "default", "node_count": 1, "deploy_type": "single_node", "env": {}, "docker_image": "vllm/vllm-openai:latest", "command": "vllm serve pfnet/plamo-3-nict-31b-base \\\n --trust-remote-code \\\n --max-model-len 4096 \\\n --tensor-parallel-size 1", "argv": [ "vllm", "serve", "pfnet/plamo-3-nict-31b-base", "--trust-remote-code", "--max-model-len", "4096", "--tensor-parallel-size", "1" ], "docker_command": "docker run --gpus all \\\n --privileged --ipc=host -p 8000:8000 \\\n -v ~/.cache/huggingface:/root/.cache/huggingface \\\n vllm/vllm-openai:latest pfnet/plamo-3-nict-31b-base \\\n --trust-remote-code \\\n --max-model-len 4096 \\\n --tensor-parallel-size 1", "docker_argv": [ "docker", "run", "--gpus", "all", "--privileged", "--ipc=host", "-p", "8000:8000", "-v", "~/.cache/huggingface:/root/.cache/huggingface", "vllm/vllm-openai:latest", "pfnet/plamo-3-nict-31b-base", "--trust-remote-code", "--max-model-len", "4096", "--tensor-parallel-size", "1" ], "strategy_spec": { "name": "single_node_tp", "deploy_type": "single_node", "display_name": "Tensor Parallel", "short_label": "TP", "orientation": "latency", "description": "Single-node tensor parallel. Splits the model across all local GPUs. TP size is set to the GPU count at deploy time. The simplest multi-GPU strategy — works for all model architectures.\n", "hardware_match": { "min_gpus": 1, "max_gpus": 8, "multi_node": false }, "vllm_args": [], "parallel_flag": "--tensor-parallel-size" }, "hardware_profile": { "brand": "NVIDIA", "generation": "hopper", "compute_arch": "sm90", "display_name": "H200", "description": "NVIDIA H200 SXM 141 GB HBM3e · 8-GPU HGX node", "gpu_count": 8, "vram_gb": 1128, "multi_node": false }, "alternatives": { "multi_node_tp": "/pfnet/plamo-3-nict-31b-base/strategies/multi_node_tp.json", "multi_node_tp_pp": "/pfnet/plamo-3-nict-31b-base/strategies/multi_node_tp_pp.json", "kv_store_centralized_mooncake": "/pfnet/plamo-3-nict-31b-base/strategies/kv_store_centralized_mooncake.json", "kv_store_distributed_mooncake": "/pfnet/plamo-3-nict-31b-base/strategies/kv_store_distributed_mooncake.json" }, "by_hardware": { "h200": "/pfnet/plamo-3-nict-31b-base/hw/h200.json", "h100": "/pfnet/plamo-3-nict-31b-base/hw/h100.json", "b200": "/pfnet/plamo-3-nict-31b-base/hw/b200.json", "gb200": "/pfnet/plamo-3-nict-31b-base/hw/gb200.json", "b300": "/pfnet/plamo-3-nict-31b-base/hw/b300.json", "gb300": "/pfnet/plamo-3-nict-31b-base/hw/gb300.json", "mi300x": "/pfnet/plamo-3-nict-31b-base/hw/mi300x.json", "mi325x": "/pfnet/plamo-3-nict-31b-base/hw/mi325x.json", "mi350x": "/pfnet/plamo-3-nict-31b-base/hw/mi350x.json", "mi355x": "/pfnet/plamo-3-nict-31b-base/hw/mi355x.json" } }, "model": { "model_id": "pfnet/plamo-3-nict-31b-base", "min_vllm_version": "0.12.0", "architecture": "dense", "parameter_count": "32B", "active_parameters": "32B", "context_length": 4096, "base_args": [ "--trust-remote-code", "--max-model-len", "4096" ], "base_env": {}, "install": { "pip": { "command": "uv venv\nsource .venv/bin/activate\nuv pip install -U vllm --torch-backend auto" }, "docker": { "command": "docker pull vllm/vllm-openai:latest" } } }, "features": {}, "opt_in_features": [], "variants": { "default": { "precision": "bf16", "vram_minimum_gb": 77, "description": "BF16 base checkpoint. Fits on H100/H200-class GPUs for 4K context; use TP=2 if extra cache headroom is needed." } }, "hardware_overrides": {}, "guide": "## Overview\n\n[PLaMo 3 NICT 31B Base](https://huggingface.co/pfnet/plamo-3-nict-31b-base)\nis the largest PLaMo 3 NICT base model from Preferred Networks and NICT.\nIt is pretrained on Japanese and English data with a hybrid attention stack\ncombining sliding-window and full-attention layers.\n\nThis is a base checkpoint. Use completion-style prompts as-is, or fine-tune\nit before deploying instruction-following or chat-style workflows.\n\n## Prerequisites\n\n- **Hardware:** H100/H200 80 GB for 4K context; use TP=2 for extra KV-cache headroom\n- **vLLM:** >= 0.12.0\n- **License:** gated Hugging Face access under the PLaMo community license\n\n### Install vLLM\n\n```bash\nuv venv\nsource .venv/bin/activate\nuv pip install -U vllm --torch-backend=auto\nhf auth login\n```\n\n## Launching the Server\n\n```bash\nvllm serve pfnet/plamo-3-nict-31b-base \\\n --trust-remote-code \\\n --max-model-len 4096\n```\n\nWith tensor parallelism for more KV-cache headroom:\n\n```bash\nvllm serve pfnet/plamo-3-nict-31b-base \\\n --trust-remote-code \\\n --max-model-len 4096 \\\n --tensor-parallel-size 2\n```\n\n## Client Usage\n\n```python\nfrom openai import OpenAI\n\nclient = OpenAI(api_key=\"EMPTY\", base_url=\"http://localhost:8000/v1\")\nresp = client.completions.create(\n model=\"pfnet/plamo-3-nict-31b-base\",\n prompt=\"The future of artificial intelligence technology is \",\n max_tokens=128,\n temperature=0.7,\n)\nprint(resp.choices[0].text)\n```\n\n## Troubleshooting\n\n- **Out of memory on 80 GB GPUs:** add `--tensor-parallel-size 2` or lower\n `--max-model-len`.\n- **Gated repo errors:** accept the PLaMo community license on Hugging Face and\n run `hf auth login`.\n- **Repetitive or low-quality completions:** adjust sampling for free-form\n generation, or fine-tune the model for chat or assistant workloads.\n\n## References\n\n- [Model card](https://huggingface.co/pfnet/plamo-3-nict-31b-base)\n- [vLLM PLaMo3 support](https://docs.vllm.ai/en/stable/models/supported_models.html)\n", "source": { "source": "vllm", "repo": "vllm-project/vllm", "ref": "main", "sha": "389ffebe555705c85337025d5df108d0384d365e", "path": "recipes/models/pfnet/plamo-3-nict-31b-base.yaml", "edit_url": "https://github.com/vllm-project/vllm/edit/main/recipes/models/pfnet/plamo-3-nict-31b-base.yaml", "overlays": [] } }