JRosenkranz's picture
updated config with base model name
03b74cd verified
raw
history blame contribute delete
448 Bytes
{
"base_model_name_or_path": "ibm-granite/granite-3b-code-instruct",
"architectures": [
"MLPSpeculatorPreTrainedModel"
],
"emb_dim": 2560,
"inner_dim": 4096,
"model_type": "mlp_speculator",
"n_candidates": 5,
"n_predict": 5,
"scale_input": false,
"tie_weights": true,
"top_k_tokens_per_head": [
4,
3,
2,
2,
2
],
"torch_dtype": "float16",
"transformers_version": "4.40.2",
"vocab_size": 49152
}