Upload folder using huggingface_hub
Browse files- README.md +18 -0
- assets/tokenizer/merges.txt +0 -0
- assets/tokenizer/vocabulary.json +0 -0
- config.json +16 -0
- metadata.json +6 -0
- model.weights.h5 +3 -0
- preprocessor.json +38 -0
- task.json +62 -0
- tokenizer.json +19 -0
README.md
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: keras-nlp
|
| 3 |
+
---
|
| 4 |
+
This is a [`Bart` model](https://keras.io/api/keras_nlp/models/bart) uploaded using the KerasNLP library and can be used with JAX, TensorFlow, and PyTorch backends.
|
| 5 |
+
This model is related to a `Seq2SeqLM` task.
|
| 6 |
+
|
| 7 |
+
Model config:
|
| 8 |
+
* **name:** bart_backbone
|
| 9 |
+
* **trainable:** True
|
| 10 |
+
* **vocabulary_size:** 50265
|
| 11 |
+
* **num_layers:** 6
|
| 12 |
+
* **num_heads:** 12
|
| 13 |
+
* **hidden_dim:** 768
|
| 14 |
+
* **intermediate_dim:** 3072
|
| 15 |
+
* **dropout:** 0.1
|
| 16 |
+
* **max_sequence_length:** 1024
|
| 17 |
+
|
| 18 |
+
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 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"module": "keras_nlp.src.models.bart.bart_backbone",
|
| 3 |
+
"class_name": "BartBackbone",
|
| 4 |
+
"config": {
|
| 5 |
+
"name": "bart_backbone",
|
| 6 |
+
"trainable": true,
|
| 7 |
+
"vocabulary_size": 50265,
|
| 8 |
+
"num_layers": 6,
|
| 9 |
+
"num_heads": 12,
|
| 10 |
+
"hidden_dim": 768,
|
| 11 |
+
"intermediate_dim": 3072,
|
| 12 |
+
"dropout": 0.1,
|
| 13 |
+
"max_sequence_length": 1024
|
| 14 |
+
},
|
| 15 |
+
"registered_name": "keras_nlp>BartBackbone"
|
| 16 |
+
}
|
metadata.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"keras_version": "3.4.1",
|
| 3 |
+
"keras_nlp_version": "0.14.0",
|
| 4 |
+
"parameter_count": 139417344,
|
| 5 |
+
"date_saved": "2024-08-13@09:54:56"
|
| 6 |
+
}
|
model.weights.h5
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:448f018e2cdfd24928c0311eb31bdb86abfe0a43f6831a987bf1ffc63723da16
|
| 3 |
+
size 558206296
|
preprocessor.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"module": "keras_nlp.src.models.bart.bart_seq_2_seq_lm_preprocessor",
|
| 3 |
+
"class_name": "BartSeq2SeqLMPreprocessor",
|
| 4 |
+
"config": {
|
| 5 |
+
"name": "bart_seq2_seq_lm_preprocessor",
|
| 6 |
+
"trainable": true,
|
| 7 |
+
"dtype": {
|
| 8 |
+
"module": "keras",
|
| 9 |
+
"class_name": "DTypePolicy",
|
| 10 |
+
"config": {
|
| 11 |
+
"name": "mixed_float16"
|
| 12 |
+
},
|
| 13 |
+
"registered_name": null
|
| 14 |
+
},
|
| 15 |
+
"tokenizer": {
|
| 16 |
+
"module": "keras_nlp.src.models.bart.bart_tokenizer",
|
| 17 |
+
"class_name": "BartTokenizer",
|
| 18 |
+
"config": {
|
| 19 |
+
"name": "bart_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 |
+
"sequence_length": null,
|
| 30 |
+
"add_prefix_space": false
|
| 31 |
+
},
|
| 32 |
+
"registered_name": "keras_nlp>BartTokenizer"
|
| 33 |
+
},
|
| 34 |
+
"encoder_sequence_length": 512,
|
| 35 |
+
"decoder_sequence_length": 128
|
| 36 |
+
},
|
| 37 |
+
"registered_name": "keras_nlp>BartSeq2SeqLMPreprocessor"
|
| 38 |
+
}
|
task.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"module": "keras_nlp.src.models.bart.bart_seq_2_seq_lm",
|
| 3 |
+
"class_name": "BartSeq2SeqLM",
|
| 4 |
+
"config": {
|
| 5 |
+
"backbone": {
|
| 6 |
+
"module": "keras_nlp.src.models.bart.bart_backbone",
|
| 7 |
+
"class_name": "BartBackbone",
|
| 8 |
+
"config": {
|
| 9 |
+
"name": "bart_backbone",
|
| 10 |
+
"trainable": true,
|
| 11 |
+
"vocabulary_size": 50265,
|
| 12 |
+
"num_layers": 6,
|
| 13 |
+
"num_heads": 12,
|
| 14 |
+
"hidden_dim": 768,
|
| 15 |
+
"intermediate_dim": 3072,
|
| 16 |
+
"dropout": 0.1,
|
| 17 |
+
"max_sequence_length": 1024
|
| 18 |
+
},
|
| 19 |
+
"registered_name": "keras_nlp>BartBackbone"
|
| 20 |
+
},
|
| 21 |
+
"preprocessor": {
|
| 22 |
+
"module": "keras_nlp.src.models.bart.bart_seq_2_seq_lm_preprocessor",
|
| 23 |
+
"class_name": "BartSeq2SeqLMPreprocessor",
|
| 24 |
+
"config": {
|
| 25 |
+
"name": "bart_seq2_seq_lm_preprocessor",
|
| 26 |
+
"trainable": true,
|
| 27 |
+
"dtype": {
|
| 28 |
+
"module": "keras",
|
| 29 |
+
"class_name": "DTypePolicy",
|
| 30 |
+
"config": {
|
| 31 |
+
"name": "mixed_float16"
|
| 32 |
+
},
|
| 33 |
+
"registered_name": null
|
| 34 |
+
},
|
| 35 |
+
"tokenizer": {
|
| 36 |
+
"module": "keras_nlp.src.models.bart.bart_tokenizer",
|
| 37 |
+
"class_name": "BartTokenizer",
|
| 38 |
+
"config": {
|
| 39 |
+
"name": "bart_tokenizer",
|
| 40 |
+
"trainable": true,
|
| 41 |
+
"dtype": {
|
| 42 |
+
"module": "keras",
|
| 43 |
+
"class_name": "DTypePolicy",
|
| 44 |
+
"config": {
|
| 45 |
+
"name": "int32"
|
| 46 |
+
},
|
| 47 |
+
"registered_name": null
|
| 48 |
+
},
|
| 49 |
+
"sequence_length": null,
|
| 50 |
+
"add_prefix_space": false
|
| 51 |
+
},
|
| 52 |
+
"registered_name": "keras_nlp>BartTokenizer"
|
| 53 |
+
},
|
| 54 |
+
"encoder_sequence_length": 512,
|
| 55 |
+
"decoder_sequence_length": 128
|
| 56 |
+
},
|
| 57 |
+
"registered_name": "keras_nlp>BartSeq2SeqLMPreprocessor"
|
| 58 |
+
},
|
| 59 |
+
"name": "bart_seq2_seq_lm"
|
| 60 |
+
},
|
| 61 |
+
"registered_name": "keras_nlp>BartSeq2SeqLM"
|
| 62 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"module": "keras_nlp.src.models.bart.bart_tokenizer",
|
| 3 |
+
"class_name": "BartTokenizer",
|
| 4 |
+
"config": {
|
| 5 |
+
"name": "bart_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 |
+
"sequence_length": null,
|
| 16 |
+
"add_prefix_space": false
|
| 17 |
+
},
|
| 18 |
+
"registered_name": "keras_nlp>BartTokenizer"
|
| 19 |
+
}
|