Training in progress, step 1000
Browse files- config.json +7 -6
- model.safetensors +2 -2
- runs/Apr05_19-42-32_9e6d5516626c/events.out.tfevents.1712346175.9e6d5516626c.440.0 +3 -0
- runs/Apr05_19-45-12_9e6d5516626c/events.out.tfevents.1712346352.9e6d5516626c.3600.0 +3 -0
- special_tokens_map.json +21 -3
- tokenizer_config.json +1 -1
- training_args.bin +1 -1
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_dropout": 0.1,
|
4 |
"activation_function": "relu",
|
5 |
"add_bias_logits": false,
|
@@ -17,13 +17,15 @@
|
|
17 |
"decoder_layerdrop": 0.0,
|
18 |
"decoder_layers": 16,
|
19 |
"decoder_start_token_id": 0,
|
|
|
20 |
"dropout": 0.1,
|
21 |
"encoder_attention_heads": 16,
|
22 |
"encoder_ffn_dim": 4096,
|
23 |
"encoder_layerdrop": 0.0,
|
24 |
"encoder_layers": 16,
|
25 |
"eos_token_id": 1,
|
26 |
-
"extra_pos_embeddings":
|
|
|
27 |
"forced_eos_token_id": 1,
|
28 |
"gradient_checkpointing": false,
|
29 |
"id2label": {
|
@@ -38,14 +40,13 @@
|
|
38 |
"LABEL_1": 1,
|
39 |
"LABEL_2": 2
|
40 |
},
|
41 |
-
"length_penalty": 0.
|
42 |
"max_length": 128,
|
43 |
-
"max_position_embeddings":
|
44 |
-
"min_length": 32,
|
45 |
"model_type": "pegasus",
|
46 |
"normalize_before": true,
|
47 |
"normalize_embedding": false,
|
48 |
-
"num_beams":
|
49 |
"num_hidden_layers": 16,
|
50 |
"pad_token_id": 0,
|
51 |
"scale_embedding": true,
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "google/pegasus-xsum",
|
3 |
"activation_dropout": 0.1,
|
4 |
"activation_function": "relu",
|
5 |
"add_bias_logits": false,
|
|
|
17 |
"decoder_layerdrop": 0.0,
|
18 |
"decoder_layers": 16,
|
19 |
"decoder_start_token_id": 0,
|
20 |
+
"do_blenderbot_90_layernorm": false,
|
21 |
"dropout": 0.1,
|
22 |
"encoder_attention_heads": 16,
|
23 |
"encoder_ffn_dim": 4096,
|
24 |
"encoder_layerdrop": 0.0,
|
25 |
"encoder_layers": 16,
|
26 |
"eos_token_id": 1,
|
27 |
+
"extra_pos_embeddings": 0,
|
28 |
+
"force_bos_token_to_be_generated": false,
|
29 |
"forced_eos_token_id": 1,
|
30 |
"gradient_checkpointing": false,
|
31 |
"id2label": {
|
|
|
40 |
"LABEL_1": 1,
|
41 |
"LABEL_2": 2
|
42 |
},
|
43 |
+
"length_penalty": 0.6,
|
44 |
"max_length": 128,
|
45 |
+
"max_position_embeddings": 512,
|
|
|
46 |
"model_type": "pegasus",
|
47 |
"normalize_before": true,
|
48 |
"normalize_embedding": false,
|
49 |
+
"num_beams": 4,
|
50 |
"num_hidden_layers": 16,
|
51 |
"pad_token_id": 0,
|
52 |
"scale_embedding": true,
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2a734ee7325e121c1aacb568aee39d37b8c11fa61bb5e1f2cc8cb84571fcb4ae
|
3 |
+
size 2279458540
|
runs/Apr05_19-42-32_9e6d5516626c/events.out.tfevents.1712346175.9e6d5516626c.440.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d6b116cb598053a67746e2ae09e3eb3fbea48f7b708a2b25486b1d86b4e35b99
|
3 |
+
size 4184
|
runs/Apr05_19-45-12_9e6d5516626c/events.out.tfevents.1712346352.9e6d5516626c.3600.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7d579c8669fc8e43eefb54bb4025a2beb6508a6cf4deaf2302dde4f06dfad4a8
|
3 |
+
size 7014
|
special_tokens_map.json
CHANGED
@@ -103,8 +103,26 @@
|
|
103 |
"<unk_101>",
|
104 |
"<unk_102>"
|
105 |
],
|
106 |
-
"eos_token":
|
|
|
|
|
|
|
|
|
|
|
|
|
107 |
"mask_token": "<mask_2>",
|
108 |
-
"pad_token":
|
109 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
110 |
}
|
|
|
103 |
"<unk_101>",
|
104 |
"<unk_102>"
|
105 |
],
|
106 |
+
"eos_token": {
|
107 |
+
"content": "</s>",
|
108 |
+
"lstrip": false,
|
109 |
+
"normalized": false,
|
110 |
+
"rstrip": false,
|
111 |
+
"single_word": false
|
112 |
+
},
|
113 |
"mask_token": "<mask_2>",
|
114 |
+
"pad_token": {
|
115 |
+
"content": "<pad>",
|
116 |
+
"lstrip": false,
|
117 |
+
"normalized": false,
|
118 |
+
"rstrip": false,
|
119 |
+
"single_word": false
|
120 |
+
},
|
121 |
+
"unk_token": {
|
122 |
+
"content": "<unk>",
|
123 |
+
"lstrip": false,
|
124 |
+
"normalized": false,
|
125 |
+
"rstrip": false,
|
126 |
+
"single_word": false
|
127 |
+
}
|
128 |
}
|
tokenizer_config.json
CHANGED
@@ -958,7 +958,7 @@
|
|
958 |
"full_tokenizer_file": null,
|
959 |
"mask_token": "<mask_2>",
|
960 |
"mask_token_sent": "<mask_1>",
|
961 |
-
"model_max_length":
|
962 |
"offset": 103,
|
963 |
"pad_token": "<pad>",
|
964 |
"sp_model_kwargs": {},
|
|
|
958 |
"full_tokenizer_file": null,
|
959 |
"mask_token": "<mask_2>",
|
960 |
"mask_token_sent": "<mask_1>",
|
961 |
+
"model_max_length": 512,
|
962 |
"offset": 103,
|
963 |
"pad_token": "<pad>",
|
964 |
"sp_model_kwargs": {},
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5112
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:943b6226bf1dc2fce9632a2428ae7481bde709ff376897a360adc4a280d61ac7
|
3 |
size 5112
|