rauma-gpt-XL / config.json
codymd's picture
Training in progress, step 500
ddf50b3 verified
raw
history blame
638 Bytes
{
"_name_or_path": "TurkuNLP/gpt3-finnish-XL",
"apply_residual_connection_post_layernorm": false,
"architectures": [
"BloomForCausalLM"
],
"attention_dropout": 0.0,
"attention_softmax_in_fp32": true,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_dropout": 0.0,
"hidden_size": 2064,
"initializer_range": 0.02,
"layer_norm_epsilon": 1e-05,
"masked_softmax_fusion": true,
"model_type": "bloom",
"n_head": 24,
"n_layer": 24,
"pad_token_id": 3,
"pretraining_tp": 1,
"slow_but_exact": false,
"torch_dtype": "float32",
"transformers_version": "4.44.2",
"use_cache": true,
"vocab_size": 131072
}