tarekziade commited on
Commit
a7f7ce0
1 Parent(s): f783bdc

New training round from scratch

Browse files
README.md CHANGED
@@ -6,7 +6,7 @@ license: apache-2.0
6
  metrics:
7
  - rouge
8
  datasets:
9
- - nlphuji/flickr30k
10
  widget:
11
  - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/savanna.jpg
12
  example_title: Savanna
@@ -24,8 +24,8 @@ model-index:
24
  type: image-to-text
25
  name: Image To Text
26
  dataset:
27
- name: nlphuji/flickr30k
28
- type: nlphuji/flickr30k
29
  metrics:
30
  - name: ROUGE-1
31
  type: rouge
@@ -61,21 +61,34 @@ This model is a work in progress. Fine-tuned version of those base models:
61
 
62
  This model was trained on:
63
 
64
- - Flickr30k : https://huggingface.co/datasets/nlphuji/flickr30k
65
- - COCO 2017: https://cocodataset.org
 
 
66
 
67
- You can get that checkpoint using the 3083a3cef6e3c8dd90df3f088074bbe836b0f403 commit.
68
 
69
- It was then further fine-tuned on :
70
 
71
- - Flickr30k debiased: https://huggingface.co/datasets/Mozilla/flickr30k-transformed-captions
72
- - DocOrNot: https://huggingface.co/datasets/Mozilla/docornot
73
 
74
- You can find the code used to create the model here: https://github.com/mozilla/distilvit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
 
76
- ### Framework versions
77
 
78
- - Transformers 4.40.2
79
- - Pytorch 2.3.0+cu121
80
- - Datasets 2.19.1
81
- - Tokenizers 0.19.1
 
6
  metrics:
7
  - rouge
8
  datasets:
9
+ - Mozilla/flickr30k-transformed-captions
10
  widget:
11
  - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/savanna.jpg
12
  example_title: Savanna
 
24
  type: image-to-text
25
  name: Image To Text
26
  dataset:
27
+ name: Mozilla/flickr30k-transformed-captions
28
+ type: Mozilla/flickr30k-transformed-captions
29
  metrics:
30
  - name: ROUGE-1
31
  type: rouge
 
61
 
62
  This model was trained on:
63
 
64
+ - [Flickr30k debiased](https://huggingface.co/datasets/Mozilla/flickr30k-transformed-captions-gpt4o)
65
+ - [DocOrNot](https://huggingface.co/datasets/Mozilla/docornot)
66
+ - [Alt Text Validation](https://huggingface.co/datasets/Mozilla/alt-text-validation)
67
+ - A debiased version of COCO 2017: https://cocodataset.org
68
 
69
+ You can find the code used to create the model here: https://github.com/mozilla/distilvit
70
 
 
71
 
72
+ # training results
 
73
 
74
+ - eval/gen_len 14.99729
75
+ - eval/loss 0.17093
76
+ - eval/meteor 0.51479
77
+ - eval/rouge1 57.8066
78
+ - eval/rouge2 35.0888
79
+ - eval/rougeL 52.9138
80
+ - eval/rougeLsum 52.9101
81
+ - eval/runtime 760.2135
82
+ - eval/samples_per_second 11.18
83
+ - eval/steps_per_second 0.112
84
+ - train/epoch 8.0
85
+ - train/global_step 11752
86
+ - train/learning_rate 0.0
87
+ - train/loss 0.1034
88
+ - train/total_flos 1.518634875573869e+20
89
+ - train/train_loss 0.14875
90
+ - train/train_runtime 91405.9053
91
+ - train/train_samples_per_second 12.855
92
+ - train/train_steps_per_second 0.129
93
 
 
94
 
 
 
 
 
config.json CHANGED
@@ -1,183 +1,183 @@
1
- {
2
- "_name_or_path": "/Volumes/SSD/vit-base-patch16-224-distilgpt2",
3
- "architectures": [
4
- "VisionEncoderDecoderModel"
5
- ],
6
- "decoder": {
7
- "_name_or_path": "distilbert/distilgpt2",
8
- "_num_labels": 1,
9
- "activation_function": "gelu_new",
10
- "add_cross_attention": true,
11
- "architectures": [
12
- "GPT2LMHeadModel"
13
- ],
14
- "attn_pdrop": 0.1,
15
- "bad_words_ids": null,
16
- "begin_suppress_tokens": null,
17
- "bos_token_id": 50256,
18
- "chunk_size_feed_forward": 0,
19
- "cross_attention_hidden_size": null,
20
- "decoder_start_token_id": null,
21
- "diversity_penalty": 0.0,
22
- "do_sample": false,
23
- "early_stopping": false,
24
- "embd_pdrop": 0.1,
25
- "encoder_no_repeat_ngram_size": 0,
26
- "eos_token_id": 50256,
27
- "exponential_decay_length_penalty": null,
28
- "finetuning_task": null,
29
- "forced_bos_token_id": null,
30
- "forced_eos_token_id": null,
31
- "id2label": {
32
- "0": "LABEL_0"
33
- },
34
- "initializer_range": 0.02,
35
- "is_decoder": true,
36
- "is_encoder_decoder": false,
37
- "label2id": {
38
- "LABEL_0": 0
39
- },
40
- "layer_norm_epsilon": 1e-05,
41
- "length_penalty": 1.0,
42
- "max_length": 20,
43
- "min_length": 0,
44
- "model_type": "gpt2",
45
- "n_ctx": 1024,
46
- "n_embd": 768,
47
- "n_head": 12,
48
- "n_inner": null,
49
- "n_layer": 6,
50
- "n_positions": 1024,
51
- "no_repeat_ngram_size": 0,
52
- "num_beam_groups": 1,
53
- "num_beams": 1,
54
- "num_return_sequences": 1,
55
- "output_attentions": false,
56
- "output_hidden_states": false,
57
- "output_scores": false,
58
- "pad_token_id": null,
59
- "prefix": null,
60
- "problem_type": null,
61
- "pruned_heads": {},
62
- "remove_invalid_values": false,
63
- "reorder_and_upcast_attn": false,
64
- "repetition_penalty": 1.0,
65
- "resid_pdrop": 0.1,
66
- "return_dict": true,
67
- "return_dict_in_generate": false,
68
- "scale_attn_by_inverse_layer_idx": false,
69
- "scale_attn_weights": true,
70
- "sep_token_id": null,
71
- "summary_activation": null,
72
- "summary_first_dropout": 0.1,
73
- "summary_proj_to_labels": true,
74
- "summary_type": "cls_index",
75
- "summary_use_proj": true,
76
- "suppress_tokens": null,
77
- "task_specific_params": {
78
- "text-generation": {
79
- "do_sample": true,
80
- "max_length": 50
81
- }
82
- },
83
- "temperature": 1.0,
84
- "tf_legacy_loss": false,
85
- "tie_encoder_decoder": false,
86
- "tie_word_embeddings": true,
87
- "tokenizer_class": null,
88
- "top_k": 50,
89
- "top_p": 1.0,
90
- "torch_dtype": null,
91
- "torchscript": false,
92
- "typical_p": 1.0,
93
- "use_bfloat16": false,
94
- "use_cache": true,
95
- "vocab_size": 50257
96
- },
97
- "decoder_start_token_id": 50256,
98
- "encoder": {
99
- "_name_or_path": "google/vit-base-patch16-224-in21k",
100
- "add_cross_attention": false,
101
- "architectures": [
102
- "ViTModel"
103
- ],
104
- "attention_probs_dropout_prob": 0.0,
105
- "bad_words_ids": null,
106
- "begin_suppress_tokens": null,
107
- "bos_token_id": null,
108
- "chunk_size_feed_forward": 0,
109
- "cross_attention_hidden_size": null,
110
- "decoder_start_token_id": null,
111
- "diversity_penalty": 0.0,
112
- "do_sample": false,
113
- "early_stopping": false,
114
- "encoder_no_repeat_ngram_size": 0,
115
- "encoder_stride": 16,
116
- "eos_token_id": null,
117
- "exponential_decay_length_penalty": null,
118
- "finetuning_task": null,
119
- "forced_bos_token_id": null,
120
- "forced_eos_token_id": null,
121
- "hidden_act": "gelu",
122
- "hidden_dropout_prob": 0.0,
123
- "hidden_size": 768,
124
- "id2label": {
125
- "0": "LABEL_0",
126
- "1": "LABEL_1"
127
- },
128
- "image_size": 224,
129
- "initializer_range": 0.02,
130
- "intermediate_size": 3072,
131
- "is_decoder": false,
132
- "is_encoder_decoder": false,
133
- "label2id": {
134
- "LABEL_0": 0,
135
- "LABEL_1": 1
136
- },
137
- "layer_norm_eps": 1e-12,
138
- "length_penalty": 1.0,
139
- "max_length": 20,
140
- "min_length": 0,
141
- "model_type": "vit",
142
- "no_repeat_ngram_size": 0,
143
- "num_attention_heads": 12,
144
- "num_beam_groups": 1,
145
- "num_beams": 1,
146
- "num_channels": 3,
147
- "num_hidden_layers": 12,
148
- "num_return_sequences": 1,
149
- "output_attentions": false,
150
- "output_hidden_states": false,
151
- "output_scores": false,
152
- "pad_token_id": null,
153
- "patch_size": 16,
154
- "prefix": null,
155
- "problem_type": null,
156
- "pruned_heads": {},
157
- "qkv_bias": true,
158
- "remove_invalid_values": false,
159
- "repetition_penalty": 1.0,
160
- "return_dict": true,
161
- "return_dict_in_generate": false,
162
- "sep_token_id": null,
163
- "suppress_tokens": null,
164
- "task_specific_params": null,
165
- "temperature": 1.0,
166
- "tf_legacy_loss": false,
167
- "tie_encoder_decoder": false,
168
- "tie_word_embeddings": true,
169
- "tokenizer_class": null,
170
- "top_k": 50,
171
- "top_p": 1.0,
172
- "torch_dtype": null,
173
- "torchscript": false,
174
- "typical_p": 1.0,
175
- "use_bfloat16": false
176
- },
177
- "eos_token_id": 50256,
178
- "is_encoder_decoder": true,
179
- "model_type": "vision-encoder-decoder",
180
- "pad_token_id": 50256,
181
- "tie_word_embeddings": false,
182
- "transformers_version": "4.33.2"
183
- }
 
1
+ {
2
+ "_name_or_path": "C:\\Users\\tarek\\Dev\\distilvit\\distilvit\\..\\vit-base-patch16-224-in21k-distilgpt2",
3
+ "architectures": [
4
+ "VisionEncoderDecoderModel"
5
+ ],
6
+ "decoder": {
7
+ "_name_or_path": "distilbert/distilgpt2",
8
+ "_num_labels": 1,
9
+ "activation_function": "gelu_new",
10
+ "add_cross_attention": true,
11
+ "architectures": [
12
+ "GPT2LMHeadModel"
13
+ ],
14
+ "attn_pdrop": 0.1,
15
+ "bad_words_ids": null,
16
+ "begin_suppress_tokens": null,
17
+ "bos_token_id": 50256,
18
+ "chunk_size_feed_forward": 0,
19
+ "cross_attention_hidden_size": null,
20
+ "decoder_start_token_id": null,
21
+ "diversity_penalty": 0.0,
22
+ "do_sample": false,
23
+ "early_stopping": false,
24
+ "embd_pdrop": 0.1,
25
+ "encoder_no_repeat_ngram_size": 0,
26
+ "eos_token_id": 50256,
27
+ "exponential_decay_length_penalty": null,
28
+ "finetuning_task": null,
29
+ "forced_bos_token_id": null,
30
+ "forced_eos_token_id": null,
31
+ "id2label": {
32
+ "0": "LABEL_0"
33
+ },
34
+ "initializer_range": 0.02,
35
+ "is_decoder": true,
36
+ "is_encoder_decoder": false,
37
+ "label2id": {
38
+ "LABEL_0": 0
39
+ },
40
+ "layer_norm_epsilon": 1e-05,
41
+ "length_penalty": 1.0,
42
+ "max_length": 20,
43
+ "min_length": 0,
44
+ "model_type": "gpt2",
45
+ "n_ctx": 1024,
46
+ "n_embd": 768,
47
+ "n_head": 12,
48
+ "n_inner": null,
49
+ "n_layer": 6,
50
+ "n_positions": 1024,
51
+ "no_repeat_ngram_size": 0,
52
+ "num_beam_groups": 1,
53
+ "num_beams": 1,
54
+ "num_return_sequences": 1,
55
+ "output_attentions": false,
56
+ "output_hidden_states": false,
57
+ "output_scores": false,
58
+ "pad_token_id": null,
59
+ "prefix": null,
60
+ "problem_type": null,
61
+ "pruned_heads": {},
62
+ "remove_invalid_values": false,
63
+ "reorder_and_upcast_attn": false,
64
+ "repetition_penalty": 1.0,
65
+ "resid_pdrop": 0.1,
66
+ "return_dict": true,
67
+ "return_dict_in_generate": false,
68
+ "scale_attn_by_inverse_layer_idx": false,
69
+ "scale_attn_weights": true,
70
+ "sep_token_id": null,
71
+ "summary_activation": null,
72
+ "summary_first_dropout": 0.1,
73
+ "summary_proj_to_labels": true,
74
+ "summary_type": "cls_index",
75
+ "summary_use_proj": true,
76
+ "suppress_tokens": null,
77
+ "task_specific_params": {
78
+ "text-generation": {
79
+ "do_sample": true,
80
+ "max_length": 50
81
+ }
82
+ },
83
+ "temperature": 1.0,
84
+ "tf_legacy_loss": false,
85
+ "tie_encoder_decoder": false,
86
+ "tie_word_embeddings": true,
87
+ "tokenizer_class": null,
88
+ "top_k": 50,
89
+ "top_p": 1.0,
90
+ "torch_dtype": null,
91
+ "torchscript": false,
92
+ "typical_p": 1.0,
93
+ "use_bfloat16": false,
94
+ "use_cache": true,
95
+ "vocab_size": 50257
96
+ },
97
+ "decoder_start_token_id": 50256,
98
+ "encoder": {
99
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
100
+ "add_cross_attention": false,
101
+ "architectures": [
102
+ "ViTModel"
103
+ ],
104
+ "attention_probs_dropout_prob": 0.0,
105
+ "bad_words_ids": null,
106
+ "begin_suppress_tokens": null,
107
+ "bos_token_id": null,
108
+ "chunk_size_feed_forward": 0,
109
+ "cross_attention_hidden_size": null,
110
+ "decoder_start_token_id": null,
111
+ "diversity_penalty": 0.0,
112
+ "do_sample": false,
113
+ "early_stopping": false,
114
+ "encoder_no_repeat_ngram_size": 0,
115
+ "encoder_stride": 16,
116
+ "eos_token_id": null,
117
+ "exponential_decay_length_penalty": null,
118
+ "finetuning_task": null,
119
+ "forced_bos_token_id": null,
120
+ "forced_eos_token_id": null,
121
+ "hidden_act": "gelu",
122
+ "hidden_dropout_prob": 0.0,
123
+ "hidden_size": 768,
124
+ "id2label": {
125
+ "0": "LABEL_0",
126
+ "1": "LABEL_1"
127
+ },
128
+ "image_size": 224,
129
+ "initializer_range": 0.02,
130
+ "intermediate_size": 3072,
131
+ "is_decoder": false,
132
+ "is_encoder_decoder": false,
133
+ "label2id": {
134
+ "LABEL_0": 0,
135
+ "LABEL_1": 1
136
+ },
137
+ "layer_norm_eps": 1e-12,
138
+ "length_penalty": 1.0,
139
+ "max_length": 20,
140
+ "min_length": 0,
141
+ "model_type": "vit",
142
+ "no_repeat_ngram_size": 0,
143
+ "num_attention_heads": 12,
144
+ "num_beam_groups": 1,
145
+ "num_beams": 1,
146
+ "num_channels": 3,
147
+ "num_hidden_layers": 12,
148
+ "num_return_sequences": 1,
149
+ "output_attentions": false,
150
+ "output_hidden_states": false,
151
+ "output_scores": false,
152
+ "pad_token_id": null,
153
+ "patch_size": 16,
154
+ "prefix": null,
155
+ "problem_type": null,
156
+ "pruned_heads": {},
157
+ "qkv_bias": true,
158
+ "remove_invalid_values": false,
159
+ "repetition_penalty": 1.0,
160
+ "return_dict": true,
161
+ "return_dict_in_generate": false,
162
+ "sep_token_id": null,
163
+ "suppress_tokens": null,
164
+ "task_specific_params": null,
165
+ "temperature": 1.0,
166
+ "tf_legacy_loss": false,
167
+ "tie_encoder_decoder": false,
168
+ "tie_word_embeddings": true,
169
+ "tokenizer_class": null,
170
+ "top_k": 50,
171
+ "top_p": 1.0,
172
+ "torch_dtype": null,
173
+ "torchscript": false,
174
+ "typical_p": 1.0,
175
+ "use_bfloat16": false
176
+ },
177
+ "eos_token_id": 50256,
178
+ "is_encoder_decoder": true,
179
+ "model_type": "vision-encoder-decoder",
180
+ "pad_token_id": 50256,
181
+ "tie_word_embeddings": false,
182
+ "transformers_version": "4.33.2"
183
+ }
generation_config.json CHANGED
@@ -1,12 +1,17 @@
1
- {
2
- "bos_token_id": 50256,
3
- "early_stopping": true,
4
- "eos_token_id": 50256,
5
- "max_length": 50,
6
- "no_repeat_ngram_size": 2,
7
- "num_beams": 2,
8
- "pad_token_id": 50256,
9
- "repetition_penalty": 1.2,
10
- "seed": 12,
11
- "transformers_version": "4.33.2"
12
- }
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 50256,
3
+ "early_stopping": true,
4
+ "eos_token_id": 50256,
5
+ "max_length": 50,
6
+ "no_repeat_ngram_size": 3,
7
+ "num_beams": 2,
8
+ "pad_token_id": 50256,
9
+ "repetition_penalty": 1.3,
10
+ "seed": 12,
11
+ "max_time": 5,
12
+ "transformers_version": "4.33.2",
13
+ "do_sample": true,
14
+ "temperature": 0.7,
15
+ "top_p": 0.9,
16
+ "top_k": 50
17
+ }
onnx/decoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e3704718aee4c999920574a3fb45b959ee59a9eb687bc5ff39c1ec6c3b3d32fc
3
  size 385864797
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74a2588b4e510214a2b4f7b89e5805e6f8d88982d2800d62ba4c3f1589957f73
3
  size 385864797
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:dfc0cac17407a66735b4b4dd5ce62900967a015a618f24c6036b628e1c600328
3
  size 387342586
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:44d84666cb213e0efdca9607f60890d21d2a6f8e74af659465b7f01ef40abda5
3
  size 387342586
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:083cc1aa2c10ee51436601f996091d4633ff2ad6ab933f1197e23c14893fc89e
3
  size 99759579
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b4f88b4343b356236aac05f47319b0f32920ff18c9ac97919dee68cca72dc8c
3
  size 99759579
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:230dd40ee4ba5bd9ec09ad783b45d98ca8acd7b37ff1f73ddbfa6a63c63163ba
3
  size 98065763
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49e086aa9b53b659ea8714e1dfd0d5ac9fa74b1a9fec949de911a43511f4fd1d
3
  size 98065763
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aae44e11c82e59d84ac7c9f654942dfbcbb2c0d2a03bc8ec49876e81dc3393a1
3
  size 385864377
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff3041032448a1af189cd6d017b5bdc82d1260481648b49b173c982ce2d17a9b
3
  size 385864377
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b1823eff0a8dc274eabb3df67551bc7330d4e22a22c23676ff978eb33c643896
3
  size 98063170
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a223f391306a0d9277b954dbad457c16a70227d1b7b8c29ce0523609c3b4468b
3
  size 98063170
onnx/encoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ca2910ada0c94d6f7d72c80e7e4f40916c0bfbd1cff763ad86d96e02fd527e90
3
  size 343440632
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f96cda5722cdd6be96aeb5f0e1901a0365bb0024936fb1feb71a107b157fda5
3
  size 343440632
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b9930375f2182d3cb829d78750939633587891b18edb0a66eeb569317ccc0776
3
- size 87038173
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:930d1c85e497cc090b1c4894bef857cf0944a0c2157d2bc5fe2d421ad290077b
3
+ size 87038170
preprocessor_config.json CHANGED
@@ -1,37 +1,23 @@
1
- {
2
- "_valid_processor_keys": [
3
- "images",
4
- "do_resize",
5
- "size",
6
- "resample",
7
- "do_rescale",
8
- "rescale_factor",
9
- "do_normalize",
10
- "image_mean",
11
- "image_std",
12
- "return_tensors",
13
- "data_format",
14
- "input_data_format"
15
- ],
16
- "do_normalize": true,
17
- "do_rescale": true,
18
- "do_resize": true,
19
- "image_mean": [
20
- 0.5,
21
- 0.5,
22
- 0.5
23
- ],
24
- "image_processor_type": "ViTFeatureExtractor",
25
- "feature_extractor_type": "ViTFeatureExtractor",
26
- "image_std": [
27
- 0.5,
28
- 0.5,
29
- 0.5
30
- ],
31
- "resample": 2,
32
- "rescale_factor": 0.00392156862745098,
33
- "size": {
34
- "height": 224,
35
- "width": 224
36
- }
37
- }
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTImageProcessor",
11
+ "feature_extractor_type": "ViTImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "resample": 2,
18
+ "rescale_factor": 0.00392156862745098,
19
+ "size": {
20
+ "height": 224,
21
+ "width": 224
22
+ }
23
+ }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c482fc553889157006b94f592050aba8b9e45475b42e8497ffa2adad938d3c7
3
+ size 730052378
quantize_config.json CHANGED
@@ -1,125 +1,125 @@
1
- {
2
- "per_channel": false,
3
- "reduce_range": false,
4
- "per_model_config": {
5
- "encoder_model": {
6
- "op_types": [
7
- "MatMul",
8
- "Add",
9
- "Equal",
10
- "ConstantOfShape",
11
- "Pow",
12
- "Erf",
13
- "Sub",
14
- "Mul",
15
- "Reshape",
16
- "Div",
17
- "Where",
18
- "Unsqueeze",
19
- "Expand",
20
- "Gather",
21
- "Softmax",
22
- "Transpose",
23
- "Constant",
24
- "Shape",
25
- "ReduceMean",
26
- "Concat",
27
- "Sqrt",
28
- "Slice",
29
- "Conv"
30
- ],
31
- "weight_type": "QUInt8"
32
- },
33
- "decoder_model": {
34
- "op_types": [
35
- "MatMul",
36
- "Add",
37
- "ConstantOfShape",
38
- "Pow",
39
- "Sub",
40
- "Mul",
41
- "Reshape",
42
- "Div",
43
- "Where",
44
- "Tanh",
45
- "Unsqueeze",
46
- "Gather",
47
- "Softmax",
48
- "Split",
49
- "Slice",
50
- "Squeeze",
51
- "Range",
52
- "Transpose",
53
- "Cast",
54
- "Constant",
55
- "Shape",
56
- "ReduceMean",
57
- "Concat",
58
- "Sqrt",
59
- "Gemm"
60
- ],
61
- "weight_type": "QInt8"
62
- },
63
- "decoder_with_past_model": {
64
- "op_types": [
65
- "MatMul",
66
- "Add",
67
- "ConstantOfShape",
68
- "Pow",
69
- "Sub",
70
- "Mul",
71
- "Reshape",
72
- "Div",
73
- "Where",
74
- "Tanh",
75
- "Unsqueeze",
76
- "Gather",
77
- "Softmax",
78
- "Split",
79
- "Slice",
80
- "Squeeze",
81
- "Range",
82
- "Transpose",
83
- "Cast",
84
- "Constant",
85
- "Shape",
86
- "ReduceMean",
87
- "Concat",
88
- "Sqrt",
89
- "Gemm"
90
- ],
91
- "weight_type": "QInt8"
92
- },
93
- "decoder_model_merged": {
94
- "op_types": [
95
- "MatMul",
96
- "Add",
97
- "ConstantOfShape",
98
- "Pow",
99
- "Sub",
100
- "Mul",
101
- "Gemm",
102
- "Reshape",
103
- "Div",
104
- "Where",
105
- "Tanh",
106
- "Unsqueeze",
107
- "Gather",
108
- "Softmax",
109
- "Split",
110
- "Squeeze",
111
- "Range",
112
- "Transpose",
113
- "Cast",
114
- "Constant",
115
- "If",
116
- "Shape",
117
- "ReduceMean",
118
- "Concat",
119
- "Sqrt",
120
- "Slice"
121
- ],
122
- "weight_type": "QInt8"
123
- }
124
- }
125
  }
 
1
+ {
2
+ "per_channel": false,
3
+ "reduce_range": false,
4
+ "per_model_config": {
5
+ "decoder_model": {
6
+ "op_types": [
7
+ "Sub",
8
+ "Concat",
9
+ "Squeeze",
10
+ "Where",
11
+ "Slice",
12
+ "Sqrt",
13
+ "Mul",
14
+ "Pow",
15
+ "Div",
16
+ "MatMul",
17
+ "Gather",
18
+ "Softmax",
19
+ "Transpose",
20
+ "ConstantOfShape",
21
+ "Reshape",
22
+ "Cast",
23
+ "Tanh",
24
+ "Unsqueeze",
25
+ "Add",
26
+ "Split",
27
+ "Shape",
28
+ "Gemm",
29
+ "Constant",
30
+ "Range",
31
+ "ReduceMean"
32
+ ],
33
+ "weight_type": "QInt8"
34
+ },
35
+ "decoder_model_merged": {
36
+ "op_types": [
37
+ "Sub",
38
+ "Concat",
39
+ "Squeeze",
40
+ "Where",
41
+ "Slice",
42
+ "Sqrt",
43
+ "Mul",
44
+ "Pow",
45
+ "Div",
46
+ "If",
47
+ "MatMul",
48
+ "Gather",
49
+ "Softmax",
50
+ "Transpose",
51
+ "ConstantOfShape",
52
+ "Reshape",
53
+ "Cast",
54
+ "Tanh",
55
+ "Unsqueeze",
56
+ "Add",
57
+ "Split",
58
+ "Shape",
59
+ "Gemm",
60
+ "Constant",
61
+ "Range",
62
+ "ReduceMean"
63
+ ],
64
+ "weight_type": "QInt8"
65
+ },
66
+ "decoder_with_past_model": {
67
+ "op_types": [
68
+ "Sub",
69
+ "Concat",
70
+ "Squeeze",
71
+ "Where",
72
+ "Slice",
73
+ "Sqrt",
74
+ "Mul",
75
+ "Pow",
76
+ "Div",
77
+ "MatMul",
78
+ "Gather",
79
+ "Softmax",
80
+ "Transpose",
81
+ "ConstantOfShape",
82
+ "Reshape",
83
+ "Cast",
84
+ "Tanh",
85
+ "Unsqueeze",
86
+ "Add",
87
+ "Split",
88
+ "Shape",
89
+ "Gemm",
90
+ "Constant",
91
+ "Range",
92
+ "ReduceMean"
93
+ ],
94
+ "weight_type": "QInt8"
95
+ },
96
+ "encoder_model": {
97
+ "op_types": [
98
+ "Sub",
99
+ "Concat",
100
+ "Where",
101
+ "Slice",
102
+ "Sqrt",
103
+ "Mul",
104
+ "Pow",
105
+ "Div",
106
+ "MatMul",
107
+ "Gather",
108
+ "Erf",
109
+ "Softmax",
110
+ "Transpose",
111
+ "Expand",
112
+ "ConstantOfShape",
113
+ "Reshape",
114
+ "Unsqueeze",
115
+ "Conv",
116
+ "Add",
117
+ "Shape",
118
+ "Equal",
119
+ "Constant",
120
+ "ReduceMean"
121
+ ],
122
+ "weight_type": "QUInt8"
123
+ }
124
+ }
125
  }
special_tokens_map.json CHANGED
@@ -1,6 +1,6 @@
1
- {
2
- "bos_token": "<|endoftext|>",
3
- "eos_token": "<|endoftext|>",
4
- "pad_token": "<|endoftext|>",
5
- "unk_token": "<|endoftext|>"
6
- }
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "pad_token": "<|endoftext|>",
5
+ "unk_token": "<|endoftext|>"
6
+ }
tokenizer_config.json CHANGED
@@ -1,20 +1,9 @@
1
- {
2
- "add_prefix_space": false,
3
- "added_tokens_decoder": {
4
- "50256": {
5
- "content": "<|endoftext|>",
6
- "lstrip": false,
7
- "normalized": true,
8
- "rstrip": false,
9
- "single_word": false,
10
- "special": true
11
- }
12
- },
13
- "bos_token": "<|endoftext|>",
14
- "clean_up_tokenization_spaces": true,
15
- "eos_token": "<|endoftext|>",
16
- "model_max_length": 1024,
17
- "pad_token": "<|endoftext|>",
18
- "tokenizer_class": "GPT2Tokenizer",
19
- "unk_token": "<|endoftext|>"
20
- }
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|endoftext|>",
6
+ "model_max_length": 1024,
7
+ "tokenizer_class": "GPT2Tokenizer",
8
+ "unk_token": "<|endoftext|>"
9
+ }
 
 
 
 
 
 
 
 
 
 
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ac92771c5176fd55689bb44c37e3467ce9f70d1a9d642ce3f6e6e4e449de33cd
3
- size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a963b881745d00ac683701b379f674fa0dbcd989df12e27bc7247cec0fdee7b
3
+ size 4664