初始化项目,由ModelHub XC社区提供模型
Model: minpeter/tiny-ko-187m-base-250725 Source: Original Platform
This commit is contained in:
30
last-checkpoint/config.json
Normal file
30
last-checkpoint/config.json
Normal file
@@ -0,0 +1,30 @@
|
||||
{
|
||||
"architectures": [
|
||||
"LlamaForCausalLM"
|
||||
],
|
||||
"attention_bias": false,
|
||||
"attention_dropout": 0.0,
|
||||
"bos_token_id": 31989,
|
||||
"eos_token_id": 31989,
|
||||
"head_dim": 64,
|
||||
"hidden_act": "silu",
|
||||
"hidden_size": 768,
|
||||
"initializer_range": 0.036084391824351615,
|
||||
"intermediate_size": 1920,
|
||||
"max_position_embeddings": 8192,
|
||||
"mlp_bias": false,
|
||||
"model_type": "llama",
|
||||
"num_attention_heads": 12,
|
||||
"num_hidden_layers": 27,
|
||||
"num_key_value_heads": 4,
|
||||
"pad_token_id": 31989,
|
||||
"pretraining_tp": 1,
|
||||
"rms_norm_eps": 1e-06,
|
||||
"rope_scaling": null,
|
||||
"rope_theta": 1000000.0,
|
||||
"tie_word_embeddings": true,
|
||||
"torch_dtype": "bfloat16",
|
||||
"transformers_version": "4.54.0",
|
||||
"use_cache": false,
|
||||
"vocab_size": 32000
|
||||
}
|
||||
8
last-checkpoint/generation_config.json
Normal file
8
last-checkpoint/generation_config.json
Normal file
@@ -0,0 +1,8 @@
|
||||
{
|
||||
"_from_model_config": true,
|
||||
"bos_token_id": 31989,
|
||||
"eos_token_id": 31989,
|
||||
"pad_token_id": 31989,
|
||||
"transformers_version": "4.54.0",
|
||||
"use_cache": false
|
||||
}
|
||||
3
last-checkpoint/model.safetensors
Normal file
3
last-checkpoint/model.safetensors
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:248a8510b4d011c6a7cb792e8c191b98f6970e721686ad59f673444ee6e8cd8d
|
||||
size 373077376
|
||||
3
last-checkpoint/optimizer.pt
Normal file
3
last-checkpoint/optimizer.pt
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:85ae475dd0fddf18ea347a8e54bf538df9aa2be9bebb564aa5efad019bafa226
|
||||
size 422378059
|
||||
3
last-checkpoint/rng_state_0.pth
Normal file
3
last-checkpoint/rng_state_0.pth
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:05104d5a3ff46ec15aca2c8ab4d432991c08c6a2122e550de6fafb4b4a7565ac
|
||||
size 14917
|
||||
3
last-checkpoint/rng_state_1.pth
Normal file
3
last-checkpoint/rng_state_1.pth
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:9741fd6cf63217d6f8cbdb641df62a9fb93f6b75fabd55cbc34ce053af823e49
|
||||
size 14917
|
||||
3
last-checkpoint/scheduler.pt
Normal file
3
last-checkpoint/scheduler.pt
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:0d3a2cef14414a6c4664da10e575c9ea9504e8280740da40bd1323bc9c67f357
|
||||
size 1401
|
||||
46135
last-checkpoint/trainer_state.json
Normal file
46135
last-checkpoint/trainer_state.json
Normal file
File diff suppressed because it is too large
Load Diff
3
last-checkpoint/training_args.bin
Normal file
3
last-checkpoint/training_args.bin
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:7907c23f07548a25486f5b0a965acb79bf0e9e3532433390bdf253cdbde7544c
|
||||
size 5777
|
||||
Reference in New Issue
Block a user