TobDeBer's picture
Upload folder using huggingface_hub
1820eac verified
{
"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"
},
"return_dict": true,
"output_hidden_states": false,
"dtype": "bfloat16",
"tie_word_embeddings": true,
"chunk_size_feed_forward": 0,
"is_encoder_decoder": false,
"is_decoder": false,
"cross_attention_hidden_size": null,
"add_cross_attention": false,
"tie_encoder_decoder": false,
"architectures": [
"LlamaForCausalLM"
],
"finetuning_task": null,
"id2label": {
"0": "LABEL_0",
"1": "LABEL_1"
},
"label2id": {
"LABEL_0": 0,
"LABEL_1": 1
},
"task_specific_params": null,
"problem_type": null,
"tokenizer_class": null,
"prefix": null,
"bos_token_id": 1,
"pad_token_id": 2,
"eos_token_id": 2,
"sep_token_id": null,
"decoder_start_token_id": null,
"_name_or_path": "./SmolLM2-135M-Instruct",
"transformers_version": "5.0.0.dev0",
"is_llama_config": true,
"model_type": "llama",
"rope_interleaved": false,
"transformers.js_config": {
"kv_cache_dtype": {
"q4f16": "float16",
"fp16": "float16"
}
},
"decomposed_layers": {
"model.layers.0.mlp.down_proj": 576,
"model.layers.0.mlp.gate_proj": 576,
"model.layers.0.mlp.up_proj": 576,
"model.layers.1.mlp.down_proj": 576,
"model.layers.1.mlp.gate_proj": 576,
"model.layers.1.mlp.up_proj": 574,
"model.layers.10.mlp.down_proj": 576,
"model.layers.10.mlp.gate_proj": 575,
"model.layers.10.mlp.up_proj": 575,
"model.layers.11.mlp.down_proj": 576,
"model.layers.11.mlp.gate_proj": 575,
"model.layers.11.mlp.up_proj": 575,
"model.layers.12.mlp.down_proj": 576,
"model.layers.12.mlp.gate_proj": 575,
"model.layers.12.mlp.up_proj": 575,
"model.layers.13.mlp.down_proj": 576,
"model.layers.13.mlp.gate_proj": 575,
"model.layers.13.mlp.up_proj": 575,
"model.layers.14.mlp.down_proj": 576,
"model.layers.14.mlp.gate_proj": 575,
"model.layers.14.mlp.up_proj": 575,
"model.layers.15.mlp.down_proj": 576,
"model.layers.15.mlp.gate_proj": 575,
"model.layers.15.mlp.up_proj": 575,
"model.layers.16.mlp.down_proj": 567,
"model.layers.16.mlp.gate_proj": 575,
"model.layers.16.mlp.up_proj": 575,
"model.layers.17.mlp.down_proj": 556,
"model.layers.17.mlp.gate_proj": 575,
"model.layers.17.mlp.up_proj": 575,
"model.layers.18.mlp.down_proj": 576,
"model.layers.18.mlp.gate_proj": 575,
"model.layers.18.mlp.up_proj": 575,
"model.layers.19.mlp.down_proj": 576,
"model.layers.19.mlp.gate_proj": 575,
"model.layers.19.mlp.up_proj": 575,
"model.layers.2.mlp.down_proj": 576,
"model.layers.2.mlp.gate_proj": 576,
"model.layers.2.mlp.up_proj": 575,
"model.layers.20.mlp.down_proj": 576,
"model.layers.20.mlp.gate_proj": 575,
"model.layers.20.mlp.up_proj": 575,
"model.layers.21.mlp.down_proj": 574,
"model.layers.21.mlp.gate_proj": 571,
"model.layers.21.mlp.up_proj": 573,
"model.layers.22.mlp.down_proj": 575,
"model.layers.22.mlp.gate_proj": 572,
"model.layers.22.mlp.up_proj": 574,
"model.layers.23.mlp.down_proj": 576,
"model.layers.23.mlp.gate_proj": 572,
"model.layers.23.mlp.up_proj": 574,
"model.layers.24.mlp.down_proj": 576,
"model.layers.24.mlp.gate_proj": 574,
"model.layers.24.mlp.up_proj": 575,
"model.layers.25.mlp.down_proj": 576,
"model.layers.25.mlp.gate_proj": 576,
"model.layers.25.mlp.up_proj": 575,
"model.layers.26.mlp.down_proj": 576,
"model.layers.26.mlp.gate_proj": 576,
"model.layers.26.mlp.up_proj": 575,
"model.layers.27.mlp.down_proj": 576,
"model.layers.27.mlp.gate_proj": 576,
"model.layers.27.mlp.up_proj": 576,
"model.layers.28.mlp.down_proj": 576,
"model.layers.28.mlp.gate_proj": 576,
"model.layers.28.mlp.up_proj": 576,
"model.layers.29.mlp.down_proj": 576,
"model.layers.29.mlp.gate_proj": 576,
"model.layers.29.mlp.up_proj": 576,
"model.layers.3.mlp.down_proj": 576,
"model.layers.3.mlp.gate_proj": 576,
"model.layers.3.mlp.up_proj": 575,
"model.layers.4.mlp.down_proj": 575,
"model.layers.4.mlp.gate_proj": 576,
"model.layers.4.mlp.up_proj": 575,
"model.layers.5.mlp.down_proj": 576,
"model.layers.5.mlp.gate_proj": 576,
"model.layers.5.mlp.up_proj": 575,
"model.layers.6.mlp.down_proj": 576,
"model.layers.6.mlp.gate_proj": 575,
"model.layers.6.mlp.up_proj": 575,
"model.layers.7.mlp.down_proj": 576,
"model.layers.7.mlp.gate_proj": 576,
"model.layers.7.mlp.up_proj": 575,
"model.layers.8.mlp.down_proj": 576,
"model.layers.8.mlp.gate_proj": 575,
"model.layers.8.mlp.up_proj": 575,
"model.layers.9.mlp.down_proj": 576,
"model.layers.9.mlp.gate_proj": 576,
"model.layers.9.mlp.up_proj": 575
},
"output_attentions": false
}