Hugging Face's logo

martin-gorner
/
ari_tiny_test_llama

Model card Files Community
martin-gorner HF staff commited on
Commit
9ebf2a8
•
1 Parent(s): fa2eb2d

Upload 8 files

Browse files
Files changed (8) hide show
  1. assets/tokenizer/merges.txt +0 -0
  2. assets/tokenizer/vocabulary.json +0 -0
  3. config.json +19 -0
  4. metadata.json +6 -0
  5. model.weights.h5 +3 -0
  6. preprocessor.json +48 -0
  7. task.json +75 -0
  8. tokenizer.json +27 -0
assets/tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
assets/tokenizer/vocabulary.json ADDED
The diff for this file is too large to render. See raw diff
 
config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_backbone",
3
+ "class_name": "Llama3Backbone",
4
+ "config": {
5
+ "name": "llama3_backbone",
6
+ "trainable": true,
7
+ "vocabulary_size": 128256,
8
+ "num_layers": 1,
9
+ "num_query_heads": 32,
10
+ "hidden_dim": 64,
11
+ "intermediate_dim": 64,
12
+ "rope_max_wavelength": 10000,
13
+ "rope_scaling_factor": 1.0,
14
+ "num_key_value_heads": 1,
15
+ "layer_norm_epsilon": 1e-06,
16
+ "dropout": 0
17
+ },
18
+ "registered_name": "keras_hub>Llama3Backbone"
19
+ }
metadata.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "keras_version": "3.6.0",
3
+ "keras_hub_version": "0.17.0.dev0",
4
+ "parameter_count": 16437696,
5
+ "date_saved": "2024-11-04@16:23:39"
6
+ }
model.weights.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02545d0637782c778f6817e5b91d731756b5adf44b37ba78b840f9c3ae184332
3
+ size 65800256
preprocessor.json ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_causal_lm_preprocessor",
3
+ "class_name": "Llama3CausalLMPreprocessor",
4
+ "config": {
5
+ "name": "llama3_causal_lm_preprocessor",
6
+ "trainable": true,
7
+ "dtype": {
8
+ "module": "keras",
9
+ "class_name": "DTypePolicy",
10
+ "config": {
11
+ "name": "float32"
12
+ },
13
+ "registered_name": null
14
+ },
15
+ "tokenizer": {
16
+ "module": "keras_hub.src.models.llama3.llama3_tokenizer",
17
+ "class_name": "Llama3Tokenizer",
18
+ "config": {
19
+ "name": "llama3_tokenizer",
20
+ "trainable": true,
21
+ "dtype": {
22
+ "module": "keras",
23
+ "class_name": "DTypePolicy",
24
+ "config": {
25
+ "name": "int32"
26
+ },
27
+ "registered_name": null
28
+ },
29
+ "config_file": "tokenizer.json",
30
+ "sequence_length": null,
31
+ "add_prefix_space": false,
32
+ "unsplittable_tokens": [
33
+ "<|end_header_id|>",
34
+ "<|end_of_text|>",
35
+ "<|eot_id|>",
36
+ "<|start_header_id|>",
37
+ "<|begin_of_text|>"
38
+ ]
39
+ },
40
+ "registered_name": "keras_hub>Llama3Tokenizer"
41
+ },
42
+ "config_file": "preprocessor.json",
43
+ "sequence_length": 1024,
44
+ "add_start_token": true,
45
+ "add_end_token": true
46
+ },
47
+ "registered_name": "keras_hub>Llama3CausalLMPreprocessor"
48
+ }
task.json ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_causal_lm",
3
+ "class_name": "Llama3CausalLM",
4
+ "config": {
5
+ "backbone": {
6
+ "module": "keras_hub.src.models.llama3.llama3_backbone",
7
+ "class_name": "Llama3Backbone",
8
+ "config": {
9
+ "name": "llama3_backbone",
10
+ "trainable": true,
11
+ "vocabulary_size": 128256,
12
+ "num_layers": 1,
13
+ "num_query_heads": 32,
14
+ "hidden_dim": 64,
15
+ "intermediate_dim": 64,
16
+ "rope_max_wavelength": 10000,
17
+ "rope_scaling_factor": 1.0,
18
+ "num_key_value_heads": 1,
19
+ "layer_norm_epsilon": 1e-06,
20
+ "dropout": 0
21
+ },
22
+ "registered_name": "keras_hub>Llama3Backbone"
23
+ },
24
+ "preprocessor": {
25
+ "module": "keras_hub.src.models.llama3.llama3_causal_lm_preprocessor",
26
+ "class_name": "Llama3CausalLMPreprocessor",
27
+ "config": {
28
+ "name": "llama3_causal_lm_preprocessor",
29
+ "trainable": true,
30
+ "dtype": {
31
+ "module": "keras",
32
+ "class_name": "DTypePolicy",
33
+ "config": {
34
+ "name": "float32"
35
+ },
36
+ "registered_name": null
37
+ },
38
+ "tokenizer": {
39
+ "module": "keras_hub.src.models.llama3.llama3_tokenizer",
40
+ "class_name": "Llama3Tokenizer",
41
+ "config": {
42
+ "name": "llama3_tokenizer",
43
+ "trainable": true,
44
+ "dtype": {
45
+ "module": "keras",
46
+ "class_name": "DTypePolicy",
47
+ "config": {
48
+ "name": "int32"
49
+ },
50
+ "registered_name": null
51
+ },
52
+ "config_file": "tokenizer.json",
53
+ "sequence_length": null,
54
+ "add_prefix_space": false,
55
+ "unsplittable_tokens": [
56
+ "<|end_header_id|>",
57
+ "<|end_of_text|>",
58
+ "<|eot_id|>",
59
+ "<|start_header_id|>",
60
+ "<|begin_of_text|>"
61
+ ]
62
+ },
63
+ "registered_name": "keras_hub>Llama3Tokenizer"
64
+ },
65
+ "config_file": "preprocessor.json",
66
+ "sequence_length": 1024,
67
+ "add_start_token": true,
68
+ "add_end_token": true
69
+ },
70
+ "registered_name": "keras_hub>Llama3CausalLMPreprocessor"
71
+ },
72
+ "name": "llama3_causal_lm"
73
+ },
74
+ "registered_name": "keras_hub>Llama3CausalLM"
75
+ }
tokenizer.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_tokenizer",
3
+ "class_name": "Llama3Tokenizer",
4
+ "config": {
5
+ "name": "llama3_tokenizer",
6
+ "trainable": true,
7
+ "dtype": {
8
+ "module": "keras",
9
+ "class_name": "DTypePolicy",
10
+ "config": {
11
+ "name": "int32"
12
+ },
13
+ "registered_name": null
14
+ },
15
+ "config_file": "tokenizer.json",
16
+ "sequence_length": null,
17
+ "add_prefix_space": false,
18
+ "unsplittable_tokens": [
19
+ "<|end_header_id|>",
20
+ "<|end_of_text|>",
21
+ "<|eot_id|>",
22
+ "<|start_header_id|>",
23
+ "<|begin_of_text|>"
24
+ ]
25
+ },
26
+ "registered_name": "keras_hub>Llama3Tokenizer"
27
+ }