QLoRA applied #2
Browse files- README.md +0 -13
- adapter_config.json +11 -7
- adapter_model.bin +2 -2
- training_args.bin +1 -1
README.md
CHANGED
@@ -2,8 +2,6 @@
|
|
2 |
base_model: ybelkada/falcon-7b-sharded-bf16
|
3 |
tags:
|
4 |
- generated_from_trainer
|
5 |
-
metrics:
|
6 |
-
- f1
|
7 |
model-index:
|
8 |
- name: falcon-7b-sharded-2
|
9 |
results: []
|
@@ -15,9 +13,6 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
# falcon-7b-sharded-2
|
16 |
|
17 |
This model is a fine-tuned version of [ybelkada/falcon-7b-sharded-bf16](https://huggingface.co/ybelkada/falcon-7b-sharded-bf16) on an unknown dataset.
|
18 |
-
It achieves the following results on the evaluation set:
|
19 |
-
- Loss: nan
|
20 |
-
- F1: 0.0337
|
21 |
|
22 |
## Model description
|
23 |
|
@@ -45,14 +40,6 @@ The following hyperparameters were used during training:
|
|
45 |
- lr_scheduler_warmup_ratio: 0.03
|
46 |
- training_steps: 500
|
47 |
|
48 |
-
### Training results
|
49 |
-
|
50 |
-
| Training Loss | Epoch | Step | Validation Loss | F1 |
|
51 |
-
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
52 |
-
| 7.6119 | 1.0 | 442 | nan | 0.0337 |
|
53 |
-
| 6.8711 | 1.13 | 500 | nan | 0.0337 |
|
54 |
-
|
55 |
-
|
56 |
### Framework versions
|
57 |
|
58 |
- Transformers 4.34.0
|
|
|
2 |
base_model: ybelkada/falcon-7b-sharded-bf16
|
3 |
tags:
|
4 |
- generated_from_trainer
|
|
|
|
|
5 |
model-index:
|
6 |
- name: falcon-7b-sharded-2
|
7 |
results: []
|
|
|
13 |
# falcon-7b-sharded-2
|
14 |
|
15 |
This model is a fine-tuned version of [ybelkada/falcon-7b-sharded-bf16](https://huggingface.co/ybelkada/falcon-7b-sharded-bf16) on an unknown dataset.
|
|
|
|
|
|
|
16 |
|
17 |
## Model description
|
18 |
|
|
|
40 |
- lr_scheduler_warmup_ratio: 0.03
|
41 |
- training_steps: 500
|
42 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
### Framework versions
|
44 |
|
45 |
- Transformers 4.34.0
|
adapter_config.json
CHANGED
@@ -1,19 +1,23 @@
|
|
1 |
{
|
|
|
2 |
"auto_mapping": null,
|
3 |
"base_model_name_or_path": "ybelkada/falcon-7b-sharded-bf16",
|
|
|
4 |
"fan_in_fan_out": false,
|
5 |
-
"feedforward_modules": [
|
6 |
-
"dense_4h_to_h",
|
7 |
-
"dense_h_to_4h"
|
8 |
-
],
|
9 |
"inference_mode": true,
|
10 |
-
"
|
|
|
|
|
|
|
|
|
11 |
"modules_to_save": null,
|
12 |
-
"peft_type": "
|
|
|
|
|
13 |
"revision": null,
|
14 |
"target_modules": [
|
15 |
-
"query_key_value",
|
16 |
"dense",
|
|
|
17 |
"dense_4h_to_h",
|
18 |
"dense_h_to_4h"
|
19 |
],
|
|
|
1 |
{
|
2 |
+
"alpha_pattern": {},
|
3 |
"auto_mapping": null,
|
4 |
"base_model_name_or_path": "ybelkada/falcon-7b-sharded-bf16",
|
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": 0.2,
|
13 |
"modules_to_save": null,
|
14 |
+
"peft_type": "LORA",
|
15 |
+
"r": 32,
|
16 |
+
"rank_pattern": {},
|
17 |
"revision": null,
|
18 |
"target_modules": [
|
|
|
19 |
"dense",
|
20 |
+
"query_key_value",
|
21 |
"dense_4h_to_h",
|
22 |
"dense_h_to_4h"
|
23 |
],
|
adapter_model.bin
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 sha256:9ea1eb39d8ceff97a117811fe16774356e662db177421214593ba8696ca1577e
|
3 |
+
size 261227285
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4091
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cc96c2a3184d35d7403313432c71e978ee6768981e0bfb651e41e3e73a264c21
|
3 |
size 4091
|