End of training
Browse files- README.md +89 -0
- adapter_config.json +28 -0
- adapter_model.safetensors +3 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- ar
|
4 |
+
library_name: peft
|
5 |
+
tags:
|
6 |
+
- generated_from_trainer
|
7 |
+
datasets:
|
8 |
+
- dalyaa/darebah2400
|
9 |
+
base_model: dalyaff/phi2-QA-Arabic-phi
|
10 |
+
model-index:
|
11 |
+
- name: phi-2
|
12 |
+
results: []
|
13 |
+
---
|
14 |
+
|
15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
16 |
+
should probably proofread and complete it, then remove this comment. -->
|
17 |
+
|
18 |
+
# phi-2
|
19 |
+
|
20 |
+
This model is a fine-tuned version of [microsoftl](https://huggingface.co/microsoftl) on the dalyaa/darebah2400 dataset.
|
21 |
+
It achieves the following results on the evaluation set:
|
22 |
+
- Loss: 0.8341
|
23 |
+
|
24 |
+
## Model description
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Intended uses & limitations
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training and evaluation data
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training procedure
|
37 |
+
|
38 |
+
### Training hyperparameters
|
39 |
+
|
40 |
+
The following hyperparameters were used during training:
|
41 |
+
- learning_rate: 2.5e-05
|
42 |
+
- train_batch_size: 2
|
43 |
+
- eval_batch_size: 8
|
44 |
+
- seed: 42
|
45 |
+
- gradient_accumulation_steps: 4
|
46 |
+
- total_train_batch_size: 8
|
47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
48 |
+
- lr_scheduler_type: linear
|
49 |
+
- lr_scheduler_warmup_steps: 5
|
50 |
+
- training_steps: 2500
|
51 |
+
|
52 |
+
### Training results
|
53 |
+
|
54 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
55 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
56 |
+
| 1.0567 | 0.4 | 100 | 1.0257 |
|
57 |
+
| 0.9463 | 0.8 | 200 | 0.9571 |
|
58 |
+
| 0.8397 | 1.2 | 300 | 0.9297 |
|
59 |
+
| 0.7876 | 1.6 | 400 | 0.9042 |
|
60 |
+
| 0.7484 | 2.0 | 500 | 0.8973 |
|
61 |
+
| 0.7301 | 2.4 | 600 | 0.8804 |
|
62 |
+
| 0.7023 | 2.8 | 700 | 0.8712 |
|
63 |
+
| 0.6604 | 3.2 | 800 | 0.8652 |
|
64 |
+
| 0.6727 | 3.6 | 900 | 0.8578 |
|
65 |
+
| 0.6542 | 4.0 | 1000 | 0.8549 |
|
66 |
+
| 0.6474 | 4.4 | 1100 | 0.8533 |
|
67 |
+
| 0.6208 | 4.8 | 1200 | 0.8503 |
|
68 |
+
| 0.6022 | 5.2 | 1300 | 0.8429 |
|
69 |
+
| 0.5997 | 5.6 | 1400 | 0.8488 |
|
70 |
+
| 0.6399 | 6.0 | 1500 | 0.8389 |
|
71 |
+
| 0.6273 | 6.4 | 1600 | 0.8410 |
|
72 |
+
| 0.5854 | 6.8 | 1700 | 0.8422 |
|
73 |
+
| 0.6062 | 7.2 | 1800 | 0.8360 |
|
74 |
+
| 0.5958 | 7.6 | 1900 | 0.8363 |
|
75 |
+
| 0.5933 | 8.0 | 2000 | 0.8403 |
|
76 |
+
| 0.5905 | 8.4 | 2100 | 0.8388 |
|
77 |
+
| 0.5604 | 8.8 | 2200 | 0.8366 |
|
78 |
+
| 0.572 | 9.2 | 2300 | 0.8349 |
|
79 |
+
| 0.5764 | 9.6 | 2400 | 0.8365 |
|
80 |
+
| 0.5926 | 10.0 | 2500 | 0.8341 |
|
81 |
+
|
82 |
+
|
83 |
+
### Framework versions
|
84 |
+
|
85 |
+
- PEFT 0.8.2
|
86 |
+
- Transformers 4.38.0.dev0
|
87 |
+
- Pytorch 2.1.0+cu121
|
88 |
+
- Datasets 2.16.1
|
89 |
+
- Tokenizers 0.15.1
|
adapter_config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "dalyaff/phi2-QA-Arabic-phi",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layers_pattern": null,
|
10 |
+
"layers_to_transform": null,
|
11 |
+
"loftq_config": {},
|
12 |
+
"lora_alpha": 16,
|
13 |
+
"lora_dropout": 0.05,
|
14 |
+
"megatron_config": null,
|
15 |
+
"megatron_core": "megatron.core",
|
16 |
+
"modules_to_save": null,
|
17 |
+
"peft_type": "LORA",
|
18 |
+
"r": 8,
|
19 |
+
"rank_pattern": {},
|
20 |
+
"revision": null,
|
21 |
+
"target_modules": [
|
22 |
+
"Wqkv",
|
23 |
+
"fc1",
|
24 |
+
"fc2"
|
25 |
+
],
|
26 |
+
"task_type": "CAUSAL_LM",
|
27 |
+
"use_rslora": false
|
28 |
+
}
|
adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a473f666daa612f997277b7ffc1c3f506b5e19cdf9d6d8901bc6599206ac7cb7
|
3 |
+
size 26230352
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a346fb380fe941fdeeb38a2f5d647775141b34e96cea23732ccb1b60ed92e9f2
|
3 |
+
size 4728
|