Hugging Face's logo

tinisoft
/
SmolLM-135M-addition

Text Generation
Transformers
Safetensors
llama
trl
sft
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
aravindhank commited on
Commit
bdb5ebe
·
verified ·
1 Parent(s): 868141e

Upload LlamaForCausalLM

Browse files
Files changed (3) hide show
  1. config.json +9 -6
  2. generation_config.json +3 -2
  3. model.safetensors +1 -1
config.json CHANGED
@@ -1,27 +1,30 @@
1
  {
2
- "_name_or_path": "HuggingFaceTB/SmolLM-135M",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
6
  "attention_bias": false,
7
  "attention_dropout": 0.0,
8
- "bos_token_id": 0,
9
- "eos_token_id": 0,
10
  "head_dim": 64,
11
  "hidden_act": "silu",
12
  "hidden_size": 576,
13
- "initializer_range": 0.02,
14
  "intermediate_size": 1536,
15
- "max_position_embeddings": 2048,
 
16
  "mlp_bias": false,
17
  "model_type": "llama",
18
  "num_attention_heads": 9,
19
  "num_hidden_layers": 30,
20
  "num_key_value_heads": 3,
 
21
  "pretraining_tp": 1,
22
  "rms_norm_eps": 1e-05,
 
23
  "rope_scaling": null,
24
- "rope_theta": 10000.0,
25
  "tie_word_embeddings": true,
26
  "torch_dtype": "float32",
27
  "transformers_version": "4.47.1",
 
1
  {
2
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
6
  "attention_bias": false,
7
  "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
  "head_dim": 64,
11
  "hidden_act": "silu",
12
  "hidden_size": 576,
13
+ "initializer_range": 0.041666666666666664,
14
  "intermediate_size": 1536,
15
+ "is_llama_config": true,
16
+ "max_position_embeddings": 8192,
17
  "mlp_bias": false,
18
  "model_type": "llama",
19
  "num_attention_heads": 9,
20
  "num_hidden_layers": 30,
21
  "num_key_value_heads": 3,
22
+ "pad_token_id": 2,
23
  "pretraining_tp": 1,
24
  "rms_norm_eps": 1e-05,
25
+ "rope_interleaved": false,
26
  "rope_scaling": null,
27
+ "rope_theta": 100000,
28
  "tie_word_embeddings": true,
29
  "torch_dtype": "float32",
30
  "transformers_version": "4.47.1",
generation_config.json CHANGED
@@ -1,6 +1,7 @@
1
  {
2
  "_from_model_config": true,
3
- "bos_token_id": 0,
4
- "eos_token_id": 0,
 
5
  "transformers_version": "4.47.1"
6
  }
 
1
  {
2
  "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 2,
6
  "transformers_version": "4.47.1"
7
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:07e4155b994bde1cadf8e6bc6a0b538b315f86cf1484d28fa086e7b6185c139c
3
  size 538090408
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cec177b89cde9983fcbc699c371032b152716b12747212b0f4e83f498d197aa6
3
  size 538090408