RefalMachine
commited on
Commit
•
47444fd
1
Parent(s):
f07d4eb
End of training
Browse files- README.md +65 -0
- adapter_config.json +25 -0
- adapter_model.bin +3 -0
- adapter_model.safetensors +3 -0
README.md
ADDED
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: RefalMachine/ruadapt_solar_10.7_part1
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
model-index:
|
6 |
+
- name: ruadapt_solar_10.7_part1_proj_init_rsg_lora
|
7 |
+
results: []
|
8 |
+
---
|
9 |
+
|
10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
11 |
+
should probably proofread and complete it, then remove this comment. -->
|
12 |
+
|
13 |
+
# ruadapt_solar_10.7_part1_proj_init_rsg_lora
|
14 |
+
|
15 |
+
This model is a fine-tuned version of [RefalMachine/ruadapt_solar_10.7_part1](https://huggingface.co/RefalMachine/ruadapt_solar_10.7_part1) on an unknown dataset.
|
16 |
+
It achieves the following results on the evaluation set:
|
17 |
+
- Loss: 0.0771
|
18 |
+
|
19 |
+
## Model description
|
20 |
+
|
21 |
+
More information needed
|
22 |
+
|
23 |
+
## Intended uses & limitations
|
24 |
+
|
25 |
+
More information needed
|
26 |
+
|
27 |
+
## Training and evaluation data
|
28 |
+
|
29 |
+
More information needed
|
30 |
+
|
31 |
+
## Training procedure
|
32 |
+
|
33 |
+
### Training hyperparameters
|
34 |
+
|
35 |
+
The following hyperparameters were used during training:
|
36 |
+
- learning_rate: 0.00025
|
37 |
+
- train_batch_size: 4
|
38 |
+
- eval_batch_size: 4
|
39 |
+
- seed: 42
|
40 |
+
- gradient_accumulation_steps: 32
|
41 |
+
- total_train_batch_size: 128
|
42 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
43 |
+
- lr_scheduler_type: cosine
|
44 |
+
- lr_scheduler_warmup_steps: 30
|
45 |
+
- num_epochs: 3
|
46 |
+
- mixed_precision_training: Native AMP
|
47 |
+
|
48 |
+
### Training results
|
49 |
+
|
50 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
51 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
52 |
+
| 0.1107 | 0.45 | 100 | 0.0784 |
|
53 |
+
| 0.0853 | 0.9 | 200 | 0.0900 |
|
54 |
+
| 0.0565 | 1.35 | 300 | 0.0818 |
|
55 |
+
| 0.0475 | 1.8 | 400 | 0.0688 |
|
56 |
+
| 0.0203 | 2.24 | 500 | 0.0788 |
|
57 |
+
| 0.0291 | 2.69 | 600 | 0.0771 |
|
58 |
+
|
59 |
+
|
60 |
+
### Framework versions
|
61 |
+
|
62 |
+
- Transformers 4.35.0
|
63 |
+
- Pytorch 2.1.2
|
64 |
+
- Datasets 2.14.4
|
65 |
+
- Tokenizers 0.14.1
|
adapter_config.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "RefalMachine/ruadapt_solar_10.7_part1",
|
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 |
+
"lora_alpha": 16,
|
12 |
+
"lora_dropout": 0.05,
|
13 |
+
"modules_to_save": null,
|
14 |
+
"peft_type": "LORA",
|
15 |
+
"r": 16,
|
16 |
+
"rank_pattern": {},
|
17 |
+
"revision": null,
|
18 |
+
"target_modules": [
|
19 |
+
"q_proj",
|
20 |
+
"o_proj",
|
21 |
+
"k_proj",
|
22 |
+
"v_proj"
|
23 |
+
],
|
24 |
+
"task_type": "CAUSAL_LM"
|
25 |
+
}
|
adapter_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ec70cfae08d7c091c122ac8dcccec1eeb5f724bace6ff432047358f0437c0083
|
3 |
+
size 81928202
|
adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d622bef4f6ebf21928c187abd70394553253194f445ea99813bc7fe5db5f4d76
|
3 |
+
size 81840624
|