TheBloke commited on
Commit
2f517c4
β€’
1 Parent(s): 4e1e1a5

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. README.md +4 -0
  2. README.mde +6 -2
  3. config.json +30 -23
  4. tokenizer.json +2 -29
  5. tokenizer.model +2 -2
README.md CHANGED
@@ -31,6 +31,10 @@ This is Transformers/HF format fp16 weights for CodeLlama 7B-Python. It is the
31
 
32
  Quantisations will be coming shortly.
33
 
 
 
 
 
34
  ## Prompt template: TBC
35
 
36
 
 
31
 
32
  Quantisations will be coming shortly.
33
 
34
+ Please note that due to a change in the RoPE Theta value, for correct results you must load these FP16 models with `trust_remote_code=True`
35
+
36
+ Credit to @emozilla for creating the necessary modelling code to achieve this!
37
+
38
  ## Prompt template: TBC
39
 
40
 
README.mde CHANGED
@@ -22,15 +22,19 @@ tags:
22
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
23
  <!-- header end -->
24
 
25
- # CodeLlama 13B-Instruct fp16
26
  - Model creator: [Meta](https://ai.meta.com/llama/)
27
 
28
  ## Description
29
 
30
- This is Transformers/HF format fp16 weights for CodeLlama 13B-Instruct. It is the result of downloading CodeLlama 13B-Instruct from [Meta](https://ai.meta.com/blog/code-llama-large-language-model-coding/) and converting to HF using `convert_llama_weights_to_hf.py`.
31
 
32
  Quantisations will be coming shortly.
33
 
 
 
 
 
34
  ## Prompt template: TBC
35
 
36
 
 
22
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
23
  <!-- header end -->
24
 
25
+ # CodeLlama %%MODEL%% fp16
26
  - Model creator: [Meta](https://ai.meta.com/llama/)
27
 
28
  ## Description
29
 
30
+ This is Transformers/HF format fp16 weights for CodeLlama %%MODEL%%. It is the result of downloading CodeLlama %%MODEL%% from [Meta](https://ai.meta.com/blog/code-llama-large-language-model-coding/) and converting to HF using `convert_llama_weights_to_hf.py`.
31
 
32
  Quantisations will be coming shortly.
33
 
34
+ Please note that due to a change in the RoPE Theta value, for correct results you must load these FP16 models with `trust_remote_code=True`
35
+
36
+ Credit to @emozilla for creating the necessary modelling code to achieve this!
37
+
38
  ## Prompt template: TBC
39
 
40
 
config.json CHANGED
@@ -1,24 +1,31 @@
1
  {
2
- "architectures": [
3
- "LlamaForCausalLM"
4
- ],
5
- "bos_token_id": 1,
6
- "eos_token_id": 2,
7
- "hidden_act": "silu",
8
- "hidden_size": 4096,
9
- "initializer_range": 0.02,
10
- "intermediate_size": 11008,
11
- "max_position_embeddings": 2048,
12
- "model_type": "llama",
13
- "num_attention_heads": 32,
14
- "num_hidden_layers": 32,
15
- "num_key_value_heads": 32,
16
- "pretraining_tp": 1,
17
- "rms_norm_eps": 1e-05,
18
- "rope_scaling": null,
19
- "tie_word_embeddings": false,
20
- "torch_dtype": "float16",
21
- "transformers_version": "4.32.0",
22
- "use_cache": true,
23
- "vocab_size": 32016
24
- }
 
 
 
 
 
 
 
 
1
  {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "bos_token_id": 1,
6
+ "eos_token_id": 2,
7
+ "hidden_act": "silu",
8
+ "hidden_size": 4096,
9
+ "initializer_range": 0.02,
10
+ "intermediate_size": 11008,
11
+ "max_position_embeddings": 16384,
12
+ "model_type": "llama",
13
+ "num_attention_heads": 32,
14
+ "num_hidden_layers": 32,
15
+ "num_key_value_heads": 32,
16
+ "pretraining_tp": 1,
17
+ "rms_norm_eps": 1e-05,
18
+ "rope_scaling": null,
19
+ "tie_word_embeddings": false,
20
+ "torch_dtype": "float16",
21
+ "transformers_version": "4.32.0",
22
+ "use_cache": true,
23
+ "vocab_size": 32016,
24
+ "auto_map": {
25
+ "AutoConfig": "configuration_llama.LlamaConfig",
26
+ "AutoModel": "modeling_llama.LlamaModel",
27
+ "AutoModelForCausalLM": "modeling_llama.LlamaForCausalLM",
28
+ "AutoModelForSequenceClassification": "modeling_llama.LlamaForSequenceClassification"
29
+ },
30
+ "rope_theta": 1000000
31
+ }
tokenizer.json CHANGED
@@ -32134,23 +32134,7 @@
32134
  "μ™•": 31996,
32135
  "ζ”Ά": 31997,
32136
  "弘": 31998,
32137
- "η»™": 31999,
32138
- "▁<SU": 32000,
32139
- "▁<SUF": 32001,
32140
- "▁<PRE": 32002,
32141
- "▁<M": 32003,
32142
- "▁<MID": 32004,
32143
- "▁<E": 32005,
32144
- "▁<EOT": 32006,
32145
- "▁<PRE>": 32007,
32146
- "▁<SUF>": 32008,
32147
- "▁<MID>": 32009,
32148
- "▁<EOT>": 32010,
32149
- "▁<EOT><EOT>": 32011,
32150
- "▁<EOT><EOT><EOT>": 32012,
32151
- "▁<EOT><EOT><EOT><EOT>": 32013,
32152
- "▁<EOT><EOT><EOT><EOT><EOT>": 32014,
32153
- "▁<EOT><EOT><EOT><EOT><EOT><EOT>": 32015
32154
  },
32155
  "merges": [
32156
  "▁ t",
@@ -93401,18 +93385,7 @@
93401
  "▁▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁",
93402
  "▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁▁",
93403
  "▁▁▁▁▁▁▁▁▁▁▁ ▁▁▁▁",
93404
- "▁ ▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
93405
- "▁< SU",
93406
- "▁<SU F",
93407
- "▁< PRE",
93408
- "▁< M",
93409
- "▁<M ID",
93410
- "▁< E",
93411
- "▁<E OT",
93412
- "▁<PRE >",
93413
- "▁<SUF >",
93414
- "▁<MID >",
93415
- "▁<EOT >"
93416
  ]
93417
  }
93418
  }
 
32134
  "μ™•": 31996,
32135
  "ζ”Ά": 31997,
32136
  "弘": 31998,
32137
+ "η»™": 31999
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
32138
  },
32139
  "merges": [
32140
  "▁ t",
 
93385
  "▁▁▁▁▁▁▁▁▁ ▁▁▁▁▁▁",
93386
  "▁▁▁▁▁▁▁ ▁▁▁▁▁▁▁▁",
93387
  "▁▁▁▁▁▁▁▁▁▁▁ ▁▁▁▁",
93388
+ "▁ ▁▁▁▁▁▁▁▁▁▁▁▁▁▁"
 
 
 
 
 
 
 
 
 
 
 
93389
  ]
93390
  }
93391
  }
tokenizer.model CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:45ccb9c8b6b561889acea59191d66986d314e7cbd6a78abc6e49b139ca91c1e6
3
- size 500058
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723