| { | |
| "architectures": [ | |
| "TinyLLMForCausalLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_tinyllm.TinyLLMConfig", | |
| "AutoModelForCausalLM": "modeling_tinyllm.TinyLLMForCausalLM" | |
| }, | |
| "hidden_size": 512, | |
| "max_position_embeddings": 512, | |
| "model_cfg": { | |
| "checkpoint_name": null, | |
| "context_window": 512, | |
| "core_model": { | |
| "attn": { | |
| "attn_type": "generic", | |
| "bias": false, | |
| "group_size": 4, | |
| "is_causal": true, | |
| "normalization": "rms_norm", | |
| "num_heads": 16 | |
| }, | |
| "ffn": { | |
| "bias": false, | |
| "ffn_dim": 1320, | |
| "ffn_type": "swiglu", | |
| "normalization": "rms_norm" | |
| }, | |
| "num_layers": 8 | |
| }, | |
| "embedding_weight_tying": true, | |
| "hidden_dim": 512, | |
| "lm_head": { | |
| "bias": false, | |
| "normalization": "rms_norm" | |
| }, | |
| "positional_encoding_type": "rope", | |
| "vocab_size": 50257 | |
| }, | |
| "model_type": "tinyllm", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 8, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.51.3", | |
| "vocab_size": 50257 | |
| } | |