CoolFace
Apppublic

sonic-coder/CPU-LLM-Inference

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
0likes
config.py148 linesDownload Raw Back to root
1# ------------------------------2# Torch-Compatible Model Definitions with Adjusted Descriptions3# ------------------------------4MODELS = {5 6    # 1B7    "K0D3IN/MiniCPM5-1B-heretic": {8        "repo_id": "K0D3IN/MiniCPM5-1B-heretic",9        "description": "K0D3IN/MiniCPM5-1B-heretic",10        "params_b": 1.011    },12    # 1.5B13    "Nemotron-Research-Reasoning-Qwen-1.5B": {14        "repo_id": "nvidia/Nemotron-Research-Reasoning-Qwen-1.5B",15        "description": "Nemotron-Research-Reasoning-Qwen-1.5B",16        "params_b": 1.517    },18    "Falcon-H1-1.5B-Instruct": {19        "repo_id": "tiiuae/Falcon-H1-1.5B-Instruct",20        "description": "Falcon‑H1 model with 1.5 B parameters, instruction‑tuned",21        "params_b": 1.522    },23    "Qwen2.5-Taiwan-1.5B-Instruct": {24        "repo_id": "benchang1110/Qwen2.5-Taiwan-1.5B-Instruct",25        "description": "Qwen2.5-Taiwan-1.5B-Instruct",26        "params_b": 1.527    },28 29    # 1.2B30    "LFM2-1.2B": {31        "repo_id": "LiquidAI/LFM2-1.2B",32        "description": "A 1.2B parameter hybrid language model from Liquid AI, designed for efficient on-device and edge AI deployment, outperforming larger models like Llama-2-7b-hf in specific tasks.",33        "params_b": 1.234    },35 36    # 1.1B37    "Taiwan-ELM-1_1B-Instruct": {38        "repo_id": "liswei/Taiwan-ELM-1_1B-Instruct",39        "description": "Taiwan-ELM-1_1B-Instruct",40        "params_b": 1.141    },42 43    # 1B44    "Llama-3.2-Taiwan-1B": {45        "repo_id": "lianghsun/Llama-3.2-Taiwan-1B",46        "description": "Llama-3.2-Taiwan base model with 1 B parameters",47        "params_b": 1.048    },49    "gemma-3-1b-it-heretic-abliterated-uncensored": {50        "repo_id": "DavidAU/gemma-3-1b-it-heretic-abliterated-uncensored",51        "description": "uncencored version of gemma-3-1b-it",52        "params_b": 1.053    },54 55    # 700M56    "LFM2-700M": {57        "repo_id": "LiquidAI/LFM2-700M",58        "description": "A 700M parameter model from the LFM2 family, designed for high efficiency on edge devices with a hybrid architecture of multiplicative gates and short convolutions.",59        "params_b": 0.760    },61 62    # 600M63    "Qwen3-0.6B": {64        "repo_id": "Qwen/Qwen3-0.6B",65        "description": "Dense causal language model with 0.6 B total parameters (0.44 B non-embedding), 28 transformer layers, 16 query heads & 8 KV heads, native 32 768-token context window, dual-mode generation, full multilingual & agentic capabilities.",66        "params_b": 0.667    },68    "Qwen3-0.6B-Taiwan": {69        "repo_id": "ShengweiPeng/Qwen3-0.6B-Taiwan",70        "description": "Qwen3-Taiwan model with 0.6 B parameters",71        "params_b": 0.672    },73 74    # 500M75    "Qwen2.5-0.5B-Taiwan-Instruct": {76        "repo_id": "ShengweiPeng/Qwen2.5-0.5B-Taiwan-Instruct",77        "description": "Qwen2.5-Taiwan model with 0.5 B parameters, instruction-tuned",78        "params_b": 0.579    },80 81    # 360M82    "SmolLM2-360M-Instruct": {83        "repo_id": "HuggingFaceTB/SmolLM2-360M-Instruct",84        "description": "Original SmolLM2‑360M Instruct",85        "params_b": 0.3686    },87    "SmolLM2-360M-Instruct-TaiwanChat": {88        "repo_id": "Luigi/SmolLM2-360M-Instruct-TaiwanChat",89        "description": "SmolLM2‑360M Instruct fine-tuned on TaiwanChat",90        "params_b": 0.3691    },92 93    # 350M94    "LFM2-350M": {95        "repo_id": "LiquidAI/LFM2-350M",96        "description": "A compact 350M parameter hybrid model optimized for edge and on-device applications, offering significantly faster training and inference speeds compared to models like Qwen3.",97        "params_b": 0.3598    },99 100    # 270M101    "parser_model_ner_gemma_v0.1": {102        "repo_id": "myfi/parser_model_ner_gemma_v0.1",103        "description": "A lightweight named‑entity‑like (NER) parser fine‑tuned from Google’s **Gemma‑3‑270M** model. The base Gemma‑3‑270M is a 270 M‑parameter, hyper‑efficient LLM designed for on‑device inference, supporting >140 languages, a 128 k‑token context window, and instruction‑following capabilities [2][7]. This variant is further trained on standard NER corpora (e.g., CoNLL‑2003, OntoNotes) to extract PERSON, ORG, LOC, and MISC entities with high precision while keeping the memory footprint low (≈240 MB VRAM in BF16 quantized form) [1]. It is released under the Apache‑2.0 license and can be used for fast, cost‑effective entity extraction in low‑resource environments.",104        "params_b": 0.27105    },106    "Gemma-3-Taiwan-270M-it": {107        "repo_id": "lianghsun/Gemma-3-Taiwan-270M-it",108        "description": "google/gemma-3-270m-it fintuned on Taiwan Chinese dataset",109        "params_b": 0.27110    },111    "gemma-3-270m-it": {112        "repo_id": "google/gemma-3-270m-it",113        "description": "Gemma‑3‑270M‑IT is a compact, 270‑million‑parameter language model fine‑tuned for Italian, offering fast and efficient on‑device text generation and comprehension in the Italian language.",114        "params_b": 0.27115    },116    "Taiwan-ELM-270M-Instruct": {117        "repo_id": "liswei/Taiwan-ELM-270M-Instruct",118        "description": "Taiwan-ELM-270M-Instruct",119        "params_b": 0.27120    },121 122    # 135M123    "SmolLM2-135M-multilingual-base": {124        "repo_id": "agentlans/SmolLM2-135M-multilingual-base",125        "description": "SmolLM2-135M-multilingual-base",126        "params_b": 0.135127    },128    "SmolLM-135M-Taiwan-Instruct-v1.0": {129        "repo_id": "benchang1110/SmolLM-135M-Taiwan-Instruct-v1.0",130        "description": "135-million-parameter F32 safetensors instruction-finetuned variant of SmolLM-135M-Taiwan, trained on the 416 k-example ChatTaiwan dataset for Traditional Chinese conversational and instruction-following tasks",131        "params_b": 0.135132    },133    "SmolLM2_135M_Grpo_Gsm8k": {134        "repo_id": "prithivMLmods/SmolLM2_135M_Grpo_Gsm8k",135        "description": "SmolLM2_135M_Grpo_Gsm8k",136        "params_b": 0.135137    },138    "SmolLM2-135M-Instruct": {139        "repo_id": "HuggingFaceTB/SmolLM2-135M-Instruct",140        "description": "Original SmolLM2‑135M Instruct",141        "params_b": 0.135142    },143    "SmolLM2-135M-Instruct-TaiwanChat": {144        "repo_id": "Luigi/SmolLM2-135M-Instruct-TaiwanChat",145        "description": "SmolLM2‑135M Instruct fine-tuned on TaiwanChat",146        "params_b": 0.135147    },148}