SREDWise commited on
Commit
fa99fc6
1 Parent(s): ce77521

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +21 -38
config.json CHANGED
@@ -1,39 +1,22 @@
1
  {
2
- "_name_or_path": "/Users/anshulachowdhury/Repo/SREDWiseAlgo/output/run_20241214_182038",
3
- "_remove_final_layer_norm": false,
4
- "activation_dropout": 0.0,
5
- "activation_function": "relu",
6
- "architectures": [
7
- "MistralForCausalLM"
8
- ],
9
- "attention_dropout": 0.0,
10
- "bos_token_id": 2,
11
- "do_layer_norm_before": true,
12
- "dropout": 0.1,
13
- "enable_bias": true,
14
- "eos_token_id": 2,
15
- "ffn_dim": 3072,
16
- "hidden_act": "silu",
17
- "hidden_size": 768,
18
- "init_std": 0.02,
19
- "initializer_range": 0.02,
20
- "intermediate_size": 14336,
21
- "layer_norm_elementwise_affine": true,
22
- "layerdrop": 0.0,
23
- "max_position_embeddings": 2048,
24
- "model_type": "mistral",
25
- "num_attention_heads": 12,
26
- "num_hidden_layers": 12,
27
- "num_key_value_heads": 8,
28
- "pad_token_id": 1,
29
- "prefix": "</s>",
30
- "rms_norm_eps": 1e-06,
31
- "rope_theta": 10000.0,
32
- "sliding_window": 4096,
33
- "tie_word_embeddings": false,
34
- "torch_dtype": "float32",
35
- "transformers_version": "4.34.0",
36
- "use_cache": true,
37
- "vocab_size": 50272,
38
- "word_embed_proj_dim": 768
39
- }
 
1
  {
2
+ "parameters": {
3
+ "model_id": "mistralai/Mistral-7B-Instruct-v0.2",
4
+ "device_map": "auto",
5
+ "torch_dtype": "bfloat16",
6
+ "max_new_tokens": 500,
7
+ "temperature": 0.7,
8
+ "top_p": 0.95,
9
+ "top_k": 50,
10
+ "num_return_sequences": 1,
11
+ "pad_token_id": 2,
12
+ "eos_token_id": 2
13
+ },
14
+ "environment": {
15
+ "CUDA_VISIBLE_DEVICES": "0",
16
+ "TRANSFORMERS_CACHE": "/cache"
17
+ },
18
+ "hardware": {
19
+ "accelerator": "gpu",
20
+ "gpu_memory": "24GB"
21
+ }
22
+ }