Hugging Face's logo

martin-gorner
/
llama-3.2-1B-pirate-instruct

Text Generation
KerasHub
Model card Files Community
martin-gorner HF staff commited on
Commit
82cbb82
•
1 Parent(s): e6a2808

Upload folder using huggingface_hub

Browse files
Files changed (9) hide show
  1. README.md +22 -0
  2. assets/tokenizer/merges.txt +0 -0
  3. assets/tokenizer/vocabulary.json +0 -0
  4. config.json +19 -0
  5. metadata.json +6 -0
  6. model.weights.h5 +3 -0
  7. preprocessor.json +51 -0
  8. task.json +78 -0
  9. tokenizer.json +30 -0
README.md ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: keras-hub
3
+ pipeline_tag: text-generation
4
+ ---
5
+ This is a [`Llama3` model](https://keras.io/api/keras_hub/models/llama3) uploaded using the KerasHub library and can be used with JAX, TensorFlow, and PyTorch backends.
6
+ This model is related to a `CausalLM` task.
7
+
8
+ Model config:
9
+ * **name:** llama3_backbone_2
10
+ * **trainable:** True
11
+ * **vocabulary_size:** 128256
12
+ * **num_layers:** 16
13
+ * **num_query_heads:** 32
14
+ * **hidden_dim:** 2048
15
+ * **intermediate_dim:** 8192
16
+ * **rope_max_wavelength:** 10000
17
+ * **rope_scaling_factor:** 1.0
18
+ * **num_key_value_heads:** 8
19
+ * **layer_norm_epsilon:** 1e-06
20
+ * **dropout:** 0
21
+
22
+ This model card has been generated automatically and should be completed by the model author. See [Model Cards documentation](https://huggingface.co/docs/hub/model-cards) for more information.
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_2",
6
+ "trainable": true,
7
+ "vocabulary_size": 128256,
8
+ "num_layers": 16,
9
+ "num_query_heads": 32,
10
+ "hidden_dim": 2048,
11
+ "intermediate_dim": 8192,
12
+ "rope_max_wavelength": 10000,
13
+ "rope_scaling_factor": 1.0,
14
+ "num_key_value_heads": 8,
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": "0.16.1",
4
+ "parameter_count": 1246316544,
5
+ "date_saved": "2024-10-18@14:08:38"
6
+ }
model.weights.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96603a1618a6cbb5efa4fdafc1f8af8a4c2182ac50c431476699a1f5ff0493b3
3
+ size 4943815248
preprocessor.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_causal_lm_preprocessor",
3
+ "class_name": "Llama3CausalLMPreprocessor",
4
+ "config": {
5
+ "name": "llama3_causal_lm_preprocessor_3",
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_4",
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
+ "<|finetune_right_pad_id|>",
34
+ "<|end_of_text|>",
35
+ "<|eom_id|>",
36
+ "<|start_header_id|>",
37
+ "<|end_header_id|>",
38
+ "<|begin_of_text|>",
39
+ "<|eot_id|>",
40
+ "<|python_tag|>"
41
+ ]
42
+ },
43
+ "registered_name": "keras_hub>Llama3Tokenizer"
44
+ },
45
+ "config_file": "preprocessor.json",
46
+ "sequence_length": 256,
47
+ "add_start_token": true,
48
+ "add_end_token": true
49
+ },
50
+ "registered_name": "keras_hub>Llama3CausalLMPreprocessor"
51
+ }
task.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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_2",
10
+ "trainable": true,
11
+ "vocabulary_size": 128256,
12
+ "num_layers": 16,
13
+ "num_query_heads": 32,
14
+ "hidden_dim": 2048,
15
+ "intermediate_dim": 8192,
16
+ "rope_max_wavelength": 10000,
17
+ "rope_scaling_factor": 1.0,
18
+ "num_key_value_heads": 8,
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_3",
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_4",
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
+ "<|finetune_right_pad_id|>",
57
+ "<|end_of_text|>",
58
+ "<|eom_id|>",
59
+ "<|start_header_id|>",
60
+ "<|end_header_id|>",
61
+ "<|begin_of_text|>",
62
+ "<|eot_id|>",
63
+ "<|python_tag|>"
64
+ ]
65
+ },
66
+ "registered_name": "keras_hub>Llama3Tokenizer"
67
+ },
68
+ "config_file": "preprocessor.json",
69
+ "sequence_length": 256,
70
+ "add_start_token": true,
71
+ "add_end_token": true
72
+ },
73
+ "registered_name": "keras_hub>Llama3CausalLMPreprocessor"
74
+ },
75
+ "name": "llama3_causal_lm_2"
76
+ },
77
+ "registered_name": "keras_hub>Llama3CausalLM"
78
+ }
tokenizer.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "module": "keras_hub.src.models.llama3.llama3_tokenizer",
3
+ "class_name": "Llama3Tokenizer",
4
+ "config": {
5
+ "name": "llama3_tokenizer_4",
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
+ "<|finetune_right_pad_id|>",
20
+ "<|end_of_text|>",
21
+ "<|eom_id|>",
22
+ "<|start_header_id|>",
23
+ "<|end_header_id|>",
24
+ "<|begin_of_text|>",
25
+ "<|eot_id|>",
26
+ "<|python_tag|>"
27
+ ]
28
+ },
29
+ "registered_name": "keras_hub>Llama3Tokenizer"
30
+ }