RadixArk/Inkling-Small-DSpark
1792
1{2 "architectures": [3 "DSparkDraftModel"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "auto_map": {8 "AutoModel": "dspark.DSparkDraftModel"9 },10 "block_size": 7,11 "bos_token_id": null,12 "confidence_head_with_markov": true,13 "dflash_config": {14 "mask_token_id": 200064,15 "target_layer_ids": [16 1,17 6,18 12,19 17,20 23,21 28,22 34,23 3924 ]25 },26 "dtype": "bfloat16",27 "enable_confidence_head": true,28 "eos_token_id": [29 200006,30 200000,31 200002,32 20000333 ],34 "head_dim": 128,35 "hidden_act": "silu",36 "hidden_size": 4096,37 "initializer_range": 0.02,38 "intermediate_size": 12288,39 "layer_types": [40 "full_attention",41 "full_attention",42 "full_attention",43 "full_attention",44 "full_attention",45 "full_attention"46 ],47 "markov_head_type": "vanilla",48 "markov_rank": 256,49 "max_position_embeddings": 1048576,50 "max_window_layers": 6,51 "model_type": "qwen3",52 "num_attention_heads": 32,53 "num_hidden_layers": 6,54 "num_key_value_heads": 8,55 "num_target_layers": 42,56 "pad_token_id": 200006,57 "rms_norm_eps": 1e-05,58 "rope_parameters": {59 "beta_fast": 32.0,60 "beta_slow": 1.0,61 "factor": 128.0,62 "original_max_position_embeddings": 8192,63 "rope_theta": 8000000,64 "rope_type": "yarn"65 },66 "rope_scaling": {67 "beta_fast": 32.0,68 "beta_slow": 1.0,69 "factor": 128.0,70 "original_max_position_embeddings": 8192,71 "rope_type": "yarn"72 },73 "rope_theta": 8000000,74 "sliding_window": null,75 "tie_word_embeddings": false,76 "transformers_version": "5.12.1",77 "use_cache": true,78 "use_sliding_window": false,79 "vocab_size": 20102480}