Text Generation
Transformers
Safetensors
7 languages
stablelm
causal-lm
Inference Endpoints
12 papers
stablelm-2-1_6b / config.json
jon-tow's picture
revert(config): use `float16` default torch dtype
a2eb1af verified
{
"architectures": [
"StableLmForCausalLM"
],
"bos_token_id": 100257,
"eos_token_id": 100257,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": 5632,
"max_position_embeddings": 4096,
"model_type": "stablelm",
"layer_norm_eps": 1e-05,
"num_attention_heads": 32,
"num_hidden_layers": 24,
"num_key_value_heads": 32,
"partial_rotary_factor": 0.25,
"rope_theta": 10000,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.38.0",
"use_cache": true,
"use_qkv_bias": true,
"vocab_size": 100352
}