Open4bits/DeepSeek-R1-mlx-2Bit
283
1{2 "architectures": [3 "DeepseekV3ForCausalLM"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "auto_map": {8 "AutoConfig": "configuration_deepseek.DeepseekV3Config",9 "AutoModel": "modeling_deepseek.DeepseekV3Model",10 "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"11 },12 "bos_token_id": 0,13 "eos_token_id": 1,14 "ep_size": 1,15 "first_k_dense_replace": 3,16 "hidden_act": "silu",17 "hidden_size": 7168,18 "initializer_range": 0.02,19 "intermediate_size": 18432,20 "kv_lora_rank": 512,21 "max_position_embeddings": 163840,22 "model_type": "deepseek_v3",23 "moe_intermediate_size": 2048,24 "moe_layer_freq": 1,25 "n_group": 8,26 "n_routed_experts": 256,27 "n_shared_experts": 1,28 "norm_topk_prob": true,29 "num_attention_heads": 128,30 "num_experts_per_tok": 8,31 "num_hidden_layers": 61,32 "num_key_value_heads": 128,33 "num_nextn_predict_layers": 1,34 "q_lora_rank": 1536,35 "qk_nope_head_dim": 128,36 "qk_rope_head_dim": 64,37 "quantization": {38 "group_size": 64,39 "bits": 2,40 "mode": "affine"41 },42 "quantization_config": {43 "group_size": 64,44 "bits": 2,45 "mode": "affine"46 },47 "rms_norm_eps": 1e-06,48 "rope_scaling": {49 "beta_fast": 32,50 "beta_slow": 1,51 "factor": 40,52 "mscale": 1.0,53 "mscale_all_dim": 1.0,54 "original_max_position_embeddings": 4096,55 "type": "yarn"56 },57 "rope_theta": 10000,58 "routed_scaling_factor": 2.5,59 "scoring_func": "sigmoid",60 "tie_word_embeddings": false,61 "topk_group": 4,62 "topk_method": "noaux_tc",63 "torch_dtype": "bfloat16",64 "transformers_version": "4.46.3",65 "use_cache": true,66 "v_head_dim": 128,67 "vocab_size": 12928068}