Simbolo commited on
Commit
38c3d14
1 Parent(s): a97e3b5

Upload model

Browse files
config.json CHANGED
@@ -1,4 +1,5 @@
1
  {
 
2
  "activation_function": "gelu_new",
3
  "architectures": [
4
  "GPT2LMHeadModel"
@@ -7,7 +8,6 @@
7
  "bos_token_id": 0,
8
  "embd_pdrop": 0.1,
9
  "eos_token_id": 5,
10
- "gradient_checkpointing": false,
11
  "initializer_range": 0.02,
12
  "layer_norm_epsilon": 1e-05,
13
  "model_type": "gpt2",
@@ -29,6 +29,6 @@
29
  "summary_use_proj": true,
30
  "torch_dtype": "float32",
31
  "transformers_version": "4.35.2",
32
- "use_cache": true,
33
  "vocab_size": 100000
34
  }
 
1
  {
2
+ "_name_or_path": "ai-forever/mGPT",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
 
8
  "bos_token_id": 0,
9
  "embd_pdrop": 0.1,
10
  "eos_token_id": 5,
 
11
  "initializer_range": 0.02,
12
  "layer_norm_epsilon": 1e-05,
13
  "model_type": "gpt2",
 
29
  "summary_use_proj": true,
30
  "torch_dtype": "float32",
31
  "transformers_version": "4.35.2",
32
+ "use_cache": false,
33
  "vocab_size": 100000
34
  }
generation_config.json CHANGED
@@ -3,5 +3,6 @@
3
  "bos_token_id": 0,
4
  "eos_token_id": 5,
5
  "pad_token_id": 1,
6
- "transformers_version": "4.35.2"
 
7
  }
 
3
  "bos_token_id": 0,
4
  "eos_token_id": 5,
5
  "pad_token_id": 1,
6
+ "transformers_version": "4.35.2",
7
+ "use_cache": false
8
  }
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3013553261f4edf7a133b53197f2bfce49b3941170f0eec5884413513ae429a4
3
  size 4998981392
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac3c658b18717ff57a88962a3991cf95c1d33e661368324922ac9d22ed686fda
3
  size 4998981392
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d7f2f0eff4079f1c5f88fef0d757212ba71b8757fcd7bae3c2cde9048953e7fa
3
  size 671436856
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6de28d6f1b32f450e225d9c823cf78a550f8c9236cb52807187d85058f1803c8
3
  size 671436856