初始化项目,由ModelHub XC社区提供模型
Model: nikitastheo/goldfish-eng-ell-sequential_interleaved Source: Original Platform
This commit is contained in:
67
training_metadata.json
Normal file
67
training_metadata.json
Normal file
@@ -0,0 +1,67 @@
|
||||
{
|
||||
"input_config": {
|
||||
"model_type": "gpt2",
|
||||
"architectures": [
|
||||
"GPT2LMHeadModel"
|
||||
],
|
||||
"activation_function": "gelu",
|
||||
"attn_pdrop": 0.1,
|
||||
"embd_pdrop": 0.1,
|
||||
"resid_pdrop": 0.1,
|
||||
"initializer_range": 0.02,
|
||||
"layer_norm_epsilon": 1e-05,
|
||||
"n_embd": 768,
|
||||
"n_inner": 3072,
|
||||
"n_head": 12,
|
||||
"n_layer": 12,
|
||||
"vocab_size": 1,
|
||||
"n_ctx": 512,
|
||||
"n_positions": 512
|
||||
},
|
||||
"num_steps_switch": 7220,
|
||||
"save_steps": [
|
||||
1,
|
||||
2,
|
||||
4,
|
||||
8,
|
||||
16,
|
||||
32,
|
||||
64,
|
||||
128,
|
||||
256,
|
||||
512,
|
||||
1000,
|
||||
2000,
|
||||
3000,
|
||||
4000,
|
||||
5000,
|
||||
6000,
|
||||
7000,
|
||||
7220,
|
||||
7221,
|
||||
7222,
|
||||
7224,
|
||||
7228,
|
||||
7236,
|
||||
7252,
|
||||
7284,
|
||||
7348,
|
||||
7476,
|
||||
7732,
|
||||
8000,
|
||||
9000,
|
||||
10000,
|
||||
11000,
|
||||
12000,
|
||||
13000,
|
||||
14000,
|
||||
15000,
|
||||
16000,
|
||||
17000,
|
||||
18000,
|
||||
19000,
|
||||
20000,
|
||||
21000,
|
||||
21640
|
||||
]
|
||||
}
|
||||
Reference in New Issue
Block a user