thanhnew2001 commited on
Commit
c6921b8
·
verified ·
1 Parent(s): 94c9e3b

Upload BloomForCausalLM

Browse files
Files changed (3) hide show
  1. config.json +5 -5
  2. generation_config.json +1 -1
  3. model.safetensors +2 -2
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "bank5",
3
  "apply_residual_connection_post_layernorm": false,
4
  "architectures": [
5
  "BloomForCausalLM"
@@ -10,14 +10,14 @@
10
  "bos_token_id": 1,
11
  "eos_token_id": 2,
12
  "hidden_dropout": 0.0,
13
- "hidden_size": 2560,
14
  "initializer_range": 0.02,
15
  "layer_norm_epsilon": 1e-05,
16
  "masked_softmax_fusion": true,
17
  "model_type": "bloom",
18
- "n_head": 32,
19
  "n_inner": null,
20
- "n_layer": 30,
21
  "offset_alibi": 100,
22
  "pad_token_id": 3,
23
  "pretraining_tp": 1,
@@ -40,7 +40,7 @@
40
  "skip_bias_add_qkv": false,
41
  "slow_but_exact": false,
42
  "torch_dtype": "float16",
43
- "transformers_version": "4.40.0.dev0",
44
  "unk_token_id": 0,
45
  "use_cache": true,
46
  "vocab_size": 250880
 
1
  {
2
+ "_name_or_path": "bank6",
3
  "apply_residual_connection_post_layernorm": false,
4
  "architectures": [
5
  "BloomForCausalLM"
 
10
  "bos_token_id": 1,
11
  "eos_token_id": 2,
12
  "hidden_dropout": 0.0,
13
+ "hidden_size": 1024,
14
  "initializer_range": 0.02,
15
  "layer_norm_epsilon": 1e-05,
16
  "masked_softmax_fusion": true,
17
  "model_type": "bloom",
18
+ "n_head": 16,
19
  "n_inner": null,
20
+ "n_layer": 24,
21
  "offset_alibi": 100,
22
  "pad_token_id": 3,
23
  "pretraining_tp": 1,
 
40
  "skip_bias_add_qkv": false,
41
  "slow_but_exact": false,
42
  "torch_dtype": "float16",
43
+ "transformers_version": "4.41.0.dev0",
44
  "unk_token_id": 0,
45
  "use_cache": true,
46
  "vocab_size": 250880
generation_config.json CHANGED
@@ -3,5 +3,5 @@
3
  "bos_token_id": 1,
4
  "eos_token_id": 2,
5
  "pad_token_id": 3,
6
- "transformers_version": "4.40.0.dev0"
7
  }
 
3
  "bos_token_id": 1,
4
  "eos_token_id": 2,
5
  "pad_token_id": 3,
6
+ "transformers_version": "4.41.0.dev0"
7
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:397a9e8f1d2b9521fccfc71dfa09b0a168d33448a204eb9b75a0b4fbdfe0f4b9
3
- size 2503879950
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0427853e136b06c2245efb69751f1a57cc09393d802450802762c7efcff0289c
3
+ size 670450446