43 lines
1.0 KiB
JSON
43 lines
1.0 KiB
JSON
{
|
|
"apply_residual_connection_post_layernorm": false,
|
|
"architectures": [
|
|
"TeleChatForCausalLM"
|
|
],
|
|
"auto_map": {
|
|
"AutoConfig": "configuration_telechat2.Telechat2Config",
|
|
"AutoModelForCausalLM": "modeling_telechat2.Telechat2ForCausalLM"
|
|
},
|
|
"attention_dropout": 0.0,
|
|
"attention_softmax_in_fp32": true,
|
|
"bias_dropout_fusion": true,
|
|
"bos_token_id": 1,
|
|
"eos_token_id": 2,
|
|
"hidden_dropout": 0.0,
|
|
"hidden_size": 4096,
|
|
"initializer_range": 0.02,
|
|
"layer_norm_epsilon": 1e-05,
|
|
"masked_softmax_fusion": true,
|
|
"model_type": "telechat",
|
|
"n_head": 32,
|
|
"num_key_value_heads":32,
|
|
"n_inner": null,
|
|
"n_layer": 30,
|
|
"offset_alibi": 100,
|
|
"pad_token_id": 3,
|
|
"pretraining_tp": 2,
|
|
"skip_bias_add": true,
|
|
"skip_bias_add_qkv": false,
|
|
"slow_but_exact": false,
|
|
"transformers_version": "4.24.0",
|
|
"unk_token_id": 0,
|
|
"use_cache": true,
|
|
"vocab_size": 131072,
|
|
"ffn_hidden_size": 12288,
|
|
"flash_attn":true,
|
|
"tie_word_embeddings":false,
|
|
"training_seqlen":8192,
|
|
"base_seqlen":8192,
|
|
"seq_length": 32768
|
|
}
|
|
|