Hugging Face's logo

ihughes15234
/
phi-3-mini-LoRA

PEFT
Safetensors
trl
sft
Generated from Trainer
Model card Files Community
ihughes15234 commited on
Commit
7e985fa
•
1 Parent(s): 74e8e04

ihughes15234/adapter-phi-3-mini-500_combined_3

Browse files
Files changed (5) hide show
  1. README.md +21 -33
  2. adapter_config.json +5 -5
  3. adapter_model.safetensors +1 -1
  4. tokenizer.json +1 -1
  5. training_args.bin +1 -1
README.md CHANGED
@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.0407
22
 
23
  ## Model description
24
 
@@ -37,7 +37,7 @@ More information needed
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
- - learning_rate: 0.0001
41
  - train_batch_size: 4
42
  - eval_batch_size: 4
43
  - seed: 42
@@ -45,42 +45,30 @@ The following hyperparameters were used during training:
45
  - total_train_batch_size: 16
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
- - lr_scheduler_warmup_ratio: 0.1
49
- - num_epochs: 5
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss |
54
  |:-------------:|:------:|:----:|:---------------:|
55
- | 1.7043 | 0.1684 | 50 | 1.6293 |
56
- | 1.2795 | 0.3367 | 100 | 0.8934 |
57
- | 0.5263 | 0.5051 | 150 | 0.1646 |
58
- | 0.0889 | 0.6734 | 200 | 0.0631 |
59
- | 0.0561 | 0.8418 | 250 | 0.0550 |
60
- | 0.0527 | 1.0101 | 300 | 0.0523 |
61
- | 0.0493 | 1.1785 | 350 | 0.0505 |
62
- | 0.048 | 1.3468 | 400 | 0.0487 |
63
- | 0.0462 | 1.5152 | 450 | 0.0478 |
64
- | 0.0482 | 1.6835 | 500 | 0.0465 |
65
- | 0.0454 | 1.8519 | 550 | 0.0455 |
66
- | 0.0443 | 2.0202 | 600 | 0.0459 |
67
- | 0.0451 | 2.1886 | 650 | 0.0455 |
68
- | 0.0447 | 2.3569 | 700 | 0.0446 |
69
- | 0.0413 | 2.5253 | 750 | 0.0441 |
70
- | 0.0423 | 2.6936 | 800 | 0.0442 |
71
- | 0.0435 | 2.8620 | 850 | 0.0437 |
72
- | 0.0411 | 3.0303 | 900 | 0.0432 |
73
- | 0.0382 | 3.1987 | 950 | 0.0431 |
74
- | 0.0413 | 3.3670 | 1000 | 0.0427 |
75
- | 0.0426 | 3.5354 | 1050 | 0.0429 |
76
- | 0.0408 | 3.7037 | 1100 | 0.0420 |
77
- | 0.0412 | 3.8721 | 1150 | 0.0418 |
78
- | 0.0377 | 4.0404 | 1200 | 0.0413 |
79
- | 0.0399 | 4.2088 | 1250 | 0.0415 |
80
- | 0.0378 | 4.3771 | 1300 | 0.0413 |
81
- | 0.037 | 4.5455 | 1350 | 0.0410 |
82
- | 0.0418 | 4.7138 | 1400 | 0.0407 |
83
- | 0.0391 | 4.8822 | 1450 | 0.0407 |
84
 
85
 
86
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.3597
22
 
23
  ## Model description
24
 
 
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
+ - learning_rate: 1e-05
41
  - train_batch_size: 4
42
  - eval_batch_size: 4
43
  - seed: 42
 
45
  - total_train_batch_size: 16
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
+ - lr_scheduler_warmup_ratio: 0.2
49
+ - num_epochs: 3
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss |
54
  |:-------------:|:------:|:----:|:---------------:|
55
+ | 1.761 | 0.1684 | 50 | 1.8291 |
56
+ | 1.751 | 0.3367 | 100 | 1.7497 |
57
+ | 1.6426 | 0.5051 | 150 | 1.6412 |
58
+ | 1.5228 | 0.6734 | 200 | 1.4889 |
59
+ | 1.3624 | 0.8418 | 250 | 1.2639 |
60
+ | 1.126 | 1.0101 | 300 | 1.0570 |
61
+ | 0.9809 | 1.1785 | 350 | 0.9361 |
62
+ | 0.8824 | 1.3468 | 400 | 0.8499 |
63
+ | 0.7793 | 1.5152 | 450 | 0.7608 |
64
+ | 0.7179 | 1.6835 | 500 | 0.6796 |
65
+ | 0.6469 | 1.8519 | 550 | 0.6057 |
66
+ | 0.5654 | 2.0202 | 600 | 0.5418 |
67
+ | 0.5096 | 2.1886 | 650 | 0.4859 |
68
+ | 0.4625 | 2.3569 | 700 | 0.4365 |
69
+ | 0.432 | 2.5253 | 750 | 0.3976 |
70
+ | 0.3866 | 2.6936 | 800 | 0.3732 |
71
+ | 0.3725 | 2.8620 | 850 | 0.3597 |
 
 
 
 
 
 
 
 
 
 
 
 
72
 
73
 
74
  ### Framework versions
adapter_config.json CHANGED
@@ -20,13 +20,13 @@
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
23
- "o_proj",
24
- "gate_proj",
25
- "down_proj",
26
  "v_proj",
27
- "up_proj",
28
  "k_proj",
29
- "q_proj"
 
 
 
 
30
  ],
31
  "task_type": "CAUSAL_LM",
32
  "use_dora": false,
 
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
 
 
 
23
  "v_proj",
 
24
  "k_proj",
25
+ "up_proj",
26
+ "q_proj",
27
+ "gate_proj",
28
+ "o_proj",
29
+ "down_proj"
30
  ],
31
  "task_type": "CAUSAL_LM",
32
  "use_dora": false,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aec11a4abbe157d1c78c298d4bf286ea5293bbb68381a8a4a815dd2c1bed1f7a
3
  size 35668592
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25b5ff67de71aad7b8738b02f01f29da452e70be0b65799e51b1f86b55789033
3
  size 35668592
tokenizer.json CHANGED
@@ -2,7 +2,7 @@
2
  "version": "1.0",
3
  "truncation": {
4
  "direction": "Right",
5
- "max_length": 1024,
6
  "strategy": "LongestFirst",
7
  "stride": 0
8
  },
 
2
  "version": "1.0",
3
  "truncation": {
4
  "direction": "Right",
5
+ "max_length": 2048,
6
  "strategy": "LongestFirst",
7
  "stride": 0
8
  },
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:05379fbffd1bbe6e3d3d99c8f29c3d067a8ceb28a47b952147c7a7356d36db87
3
  size 5432
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca4861df2247dd78029c41574a03528f12104d588aabeff4c3b6f35434318fa0
3
  size 5432