Hugging Face's logo

Ro551
/
opus-mt-es-en-GEC-spanish-LORA-cowsl2h

PEFT
Safetensors
Transformers
lora
Model card Files
xet
 Community
Ro551 commited on
Commit
3d1968d
·
verified ·
1 Parent(s): 24dfbbd

Model trained and pushed to Hugging Face Hub

Browse files
Files changed (4) hide show
  1. README.md +9 -8
  2. adapter_config.json +4 -4
  3. adapter_model.safetensors +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [Helsinki-NLP/opus-mt-es-en](https://huggingface.co/Helsinki-NLP/opus-mt-es-en) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.1722
22
- - Gleu: 0.3735
23
 
24
  ## Model description
25
 
@@ -42,21 +42,22 @@ The following hyperparameters were used during training:
42
  - train_batch_size: 8
43
  - eval_batch_size: 8
44
  - seed: 42
45
- - gradient_accumulation_steps: 8
46
- - total_train_batch_size: 64
47
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
48
  - lr_scheduler_type: linear
49
  - lr_scheduler_warmup_steps: 0.09857017609527079
50
- - num_epochs: 3
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Gleu |
56
  |:-------------:|:-----:|:----:|:---------------:|:------:|
57
- | No log | 1.0 | 97 | 0.2185 | 0.3091 |
58
- | 9.8620 | 2.0 | 194 | 0.1818 | 0.3603 |
59
- | 1.6898 | 3.0 | 291 | 0.1722 | 0.3735 |
 
60
 
61
 
62
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [Helsinki-NLP/opus-mt-es-en](https://huggingface.co/Helsinki-NLP/opus-mt-es-en) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.1340
22
+ - Gleu: 0.4355
23
 
24
  ## Model description
25
 
 
42
  - train_batch_size: 8
43
  - eval_batch_size: 8
44
  - seed: 42
45
+ - gradient_accumulation_steps: 2
46
+ - total_train_batch_size: 16
47
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
48
  - lr_scheduler_type: linear
49
  - lr_scheduler_warmup_steps: 0.09857017609527079
50
+ - num_epochs: 4
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Gleu |
56
  |:-------------:|:-----:|:----:|:---------------:|:------:|
57
+ | 0.4962 | 1.0 | 388 | 0.1879 | 0.3553 |
58
+ | 0.3441 | 2.0 | 776 | 0.1530 | 0.3935 |
59
+ | 0.2641 | 3.0 | 1164 | 0.1392 | 0.4264 |
60
+ | 0.2190 | 4.0 | 1552 | 0.1340 | 0.4355 |
61
 
62
 
63
  ### Framework versions
adapter_config.json CHANGED
@@ -29,12 +29,12 @@
29
  "rank_pattern": {},
30
  "revision": null,
31
  "target_modules": [
32
- "k_proj",
33
  "fc1",
 
34
  "v_proj",
35
- "out_proj",
36
- "q_proj",
37
- "fc2"
38
  ],
39
  "target_parameters": null,
40
  "task_type": "SEQ_2_SEQ_LM",
 
29
  "rank_pattern": {},
30
  "revision": null,
31
  "target_modules": [
32
+ "q_proj",
33
  "fc1",
34
+ "fc2",
35
  "v_proj",
36
+ "k_proj",
37
+ "out_proj"
 
38
  ],
39
  "target_parameters": null,
40
  "task_type": "SEQ_2_SEQ_LM",
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:afa13c862536f67747d3e0b88a8b8472f56654b249a28e368736d53fc98b41f4
3
  size 17327976
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15504a67ce276ed5be80757b437c187f4bfb29e47408a024e9f48a093be355a4
3
  size 17327976
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:029fe4d67eeb01f41180aa10880fbc747bbe73c0684149bc5609442033fa8523
3
  size 5393
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e265df70fbd3b4428715f75a424de20588ff58bd81ecb900f1276c2e52259cc
3
  size 5393