numen-tech commited on
Commit
ae30531
1 Parent(s): a858427

Update model

Browse files
Files changed (3) hide show
  1. ndarray-cache.json +1 -1
  2. private-llm-config.json +5 -3
  3. tokenizer.json +0 -0
ndarray-cache.json CHANGED
@@ -2,7 +2,7 @@
2
  "metadata": {
3
  "ParamSize": 325,
4
  "ParamBytes": 3734421760.0,
5
- "BitsPerParam": 4.125436579647682
6
  },
7
  "records": [
8
  {
 
2
  "metadata": {
3
  "ParamSize": 325,
4
  "ParamBytes": 3734421760.0,
5
+ "BitsPerParam": 3.0339086581997106
6
  },
7
  "records": [
8
  {
private-llm-config.json CHANGED
@@ -22,21 +22,23 @@
22
  "prefill_chunk_size": 128,
23
  "attention_sink_size": 4,
24
  "tensor_parallel_shards": 1,
25
- "max_batch_size": 80,
26
  "mean_gen_len": 128,
27
  "max_gen_len": 512,
28
  "shift_fill_factor": 0.3,
29
  "temperature": 0.7,
 
 
30
  "repetition_penalty": 1.0,
31
  "top_p": 0.95,
32
- "conv_template": "openchat3.5",
33
  "pad_token_id": 0,
34
  "bos_token_id": 1,
35
  "eos_token_id": 2,
36
  "tokenizer_files": [
37
  "tokenizer.model",
38
  "added_tokens.json",
39
- "tokenizer_config.json"
 
40
  ],
41
  "version": "0.1.0"
42
  }
 
22
  "prefill_chunk_size": 128,
23
  "attention_sink_size": 4,
24
  "tensor_parallel_shards": 1,
 
25
  "mean_gen_len": 128,
26
  "max_gen_len": 512,
27
  "shift_fill_factor": 0.3,
28
  "temperature": 0.7,
29
+ "presence_penalty": 0.0,
30
+ "frequency_penalty": 0.0,
31
  "repetition_penalty": 1.0,
32
  "top_p": 0.95,
33
+ "conv_template": "mistral_default",
34
  "pad_token_id": 0,
35
  "bos_token_id": 1,
36
  "eos_token_id": 2,
37
  "tokenizer_files": [
38
  "tokenizer.model",
39
  "added_tokens.json",
40
+ "tokenizer_config.json",
41
+ "tokenizer.json"
42
  ],
43
  "version": "0.1.0"
44
  }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff