{ "vocab_size": 49152, "max_position_embeddings": 8192, "hidden_size": 576, "intermediate_size": 1536, "num_hidden_layers": 30, "num_attention_heads": 9, "num_key_value_heads": 3, "hidden_act": "silu", "initializer_range": 0.041666666666666664, "rms_norm_eps": 1e-05, "pretraining_tp": 1, "use_cache": true, "attention_bias": false, "attention_dropout": 0.0, "mlp_bias": false, "head_dim": 64, "rope_parameters": { "rope_theta": 100000, "rope_type": "default" }, "tie_word_embeddings": true, "pad_token_id": 2, "bos_token_id": 1, "eos_token_id": 2, "return_dict": true, "output_hidden_states": false, "dtype": "bfloat16", "chunk_size_feed_forward": 0, "is_encoder_decoder": false, "architectures": [ "LlamaForCausalLM" ], "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "problem_type": null, "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct", "transformers_version": "5.3.0", "is_llama_config": true, "model_type": "llama", "rope_interleaved": false, "transformers.js_config": { "kv_cache_dtype": { "q4f16": "float16", "fp16": "float16" } }, "output_attentions": false, "quantization_config": { "quant_method": "glq", "codebook": "e8_shell", "codesz": 8, "bpw": 4 } }