sri-lasya/gst-taxing-llm
Browse files- README.md +12 -27
- adapter_config.json +3 -3
- adapter_model.safetensors +1 -1
- training_args.bin +1 -1
README.md
CHANGED
@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
17 |
|
18 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 1.
|
21 |
|
22 |
## Model description
|
23 |
|
@@ -43,37 +43,22 @@ The following hyperparameters were used during training:
|
|
43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
44 |
- lr_scheduler_type: constant
|
45 |
- lr_scheduler_warmup_steps: 0.03
|
46 |
-
- training_steps:
|
47 |
|
48 |
### Training results
|
49 |
|
50 |
| Training Loss | Epoch | Step | Validation Loss |
|
51 |
|:-------------:|:-----:|:----:|:---------------:|
|
52 |
-
| 3.
|
53 |
-
| 2.
|
54 |
-
| 1.
|
55 |
-
| 1.
|
56 |
-
| 1.
|
57 |
-
| 1.
|
58 |
-
| 1.
|
59 |
-
| 1.
|
60 |
-
| 1.
|
61 |
-
| 1.
|
62 |
-
| 1.7315 | 0.44 | 110 | 1.7449 |
|
63 |
-
| 1.2828 | 0.48 | 120 | 1.7161 |
|
64 |
-
| 1.5608 | 0.52 | 130 | 1.6869 |
|
65 |
-
| 1.5608 | 0.56 | 140 | 1.6915 |
|
66 |
-
| 1.4046 | 0.6 | 150 | 1.6782 |
|
67 |
-
| 1.3853 | 0.64 | 160 | 1.6715 |
|
68 |
-
| 1.4398 | 0.68 | 170 | 1.6472 |
|
69 |
-
| 1.5917 | 0.72 | 180 | 1.6333 |
|
70 |
-
| 1.7207 | 0.76 | 190 | 1.6295 |
|
71 |
-
| 1.5135 | 0.8 | 200 | 1.6901 |
|
72 |
-
| 1.3951 | 0.84 | 210 | 1.7478 |
|
73 |
-
| 1.6068 | 0.88 | 220 | 1.7256 |
|
74 |
-
| 1.7094 | 0.92 | 230 | 1.7206 |
|
75 |
-
| 1.8219 | 0.96 | 240 | 1.7528 |
|
76 |
-
| 1.7888 | 1.0 | 250 | 1.7859 |
|
77 |
|
78 |
|
79 |
### Framework versions
|
|
|
17 |
|
18 |
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 1.7777
|
21 |
|
22 |
## Model description
|
23 |
|
|
|
43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
44 |
- lr_scheduler_type: constant
|
45 |
- lr_scheduler_warmup_steps: 0.03
|
46 |
+
- training_steps: 100
|
47 |
|
48 |
### Training results
|
49 |
|
50 |
| Training Loss | Epoch | Step | Validation Loss |
|
51 |
|:-------------:|:-----:|:----:|:---------------:|
|
52 |
+
| 3.1448 | 0.1 | 10 | 2.6930 |
|
53 |
+
| 2.5114 | 0.2 | 20 | 2.1804 |
|
54 |
+
| 1.6951 | 0.3 | 30 | 2.0022 |
|
55 |
+
| 1.7202 | 0.4 | 40 | 1.9292 |
|
56 |
+
| 1.7303 | 0.5 | 50 | 1.9083 |
|
57 |
+
| 1.4861 | 0.6 | 60 | 1.8414 |
|
58 |
+
| 1.7349 | 0.7 | 70 | 1.7941 |
|
59 |
+
| 1.6208 | 0.8 | 80 | 1.7469 |
|
60 |
+
| 1.6669 | 0.9 | 90 | 1.7685 |
|
61 |
+
| 1.5728 | 1.0 | 100 | 1.7777 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
62 |
|
63 |
|
64 |
### Framework versions
|
adapter_config.json
CHANGED
@@ -21,11 +21,11 @@
|
|
21 |
"revision": null,
|
22 |
"target_modules": [
|
23 |
"k_proj",
|
24 |
-
"
|
25 |
-
"gate_proj",
|
26 |
"v_proj",
|
27 |
"q_proj",
|
28 |
-
"
|
|
|
29 |
"o_proj",
|
30 |
"lm_head"
|
31 |
],
|
|
|
21 |
"revision": null,
|
22 |
"target_modules": [
|
23 |
"k_proj",
|
24 |
+
"up_proj",
|
|
|
25 |
"v_proj",
|
26 |
"q_proj",
|
27 |
+
"down_proj",
|
28 |
+
"gate_proj",
|
29 |
"o_proj",
|
30 |
"lm_head"
|
31 |
],
|
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1217458040
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:882aa51c16706785857813f589e45856756f4e694354b248a3bbd8616a6d5a7d
|
3 |
size 1217458040
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5048
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5c4bcac94e7d8a069e668bc9663b1ba2feffdb9f1ba69641a598006ebedc3773
|
3 |
size 5048
|