dpfried commited on
Commit
bbb9990
1 Parent(s): c656d90
Files changed (2) hide show
  1. config.json +27 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/checkpoint/dpf/models/cm-1.3B-hf",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "XGLMForCausalLM"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "attention_heads": 32,
10
+ "bos_token_id": 0,
11
+ "d_model": 2048,
12
+ "decoder_start_token_id": 2,
13
+ "dropout": 0.1,
14
+ "eos_token_id": 2,
15
+ "ffn_dim": 8192,
16
+ "init_std": 0.02,
17
+ "layerdrop": 0.0,
18
+ "max_position_embeddings": 2048,
19
+ "model_type": "xglm",
20
+ "num_layers": 24,
21
+ "pad_token_id": 1,
22
+ "scale_embedding": true,
23
+ "torch_dtype": "float16",
24
+ "transformers_version": "4.18.0.dev0",
25
+ "use_cache": true,
26
+ "vocab_size": 50518
27
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a38f8f491129fa23d24aa74df39ad02ba172fd6332d20714580b0496c240ac8
3
+ size 2624252219