Commit
•
93f3dec
1
Parent(s):
c6fda03
lucifer/qlora-news
Browse files
README.md
CHANGED
|
@@ -17,8 +17,6 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
# outputdir
|
| 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: 1.2794
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -48,37 +46,6 @@ The following hyperparameters were used during training:
|
|
| 48 |
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
- num_epochs: 2
|
| 50 |
|
| 51 |
-
### Training results
|
| 52 |
-
|
| 53 |
-
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
-
|:-------------:|:------:|:----:|:---------------:|
|
| 55 |
-
| 3.0506 | 0.0771 | 100 | 2.8512 |
|
| 56 |
-
| 2.4484 | 0.1542 | 200 | 1.7779 |
|
| 57 |
-
| 1.5032 | 0.2313 | 300 | 1.3608 |
|
| 58 |
-
| 1.3472 | 0.3085 | 400 | 1.3325 |
|
| 59 |
-
| 1.3304 | 0.3856 | 500 | 1.3195 |
|
| 60 |
-
| 1.3156 | 0.4627 | 600 | 1.3103 |
|
| 61 |
-
| 1.313 | 0.5398 | 700 | 1.3038 |
|
| 62 |
-
| 1.2976 | 0.6169 | 800 | 1.2991 |
|
| 63 |
-
| 1.3001 | 0.6940 | 900 | 1.2956 |
|
| 64 |
-
| 1.2976 | 0.7712 | 1000 | 1.2927 |
|
| 65 |
-
| 1.2902 | 0.8483 | 1100 | 1.2907 |
|
| 66 |
-
| 1.2831 | 0.9254 | 1200 | 1.2888 |
|
| 67 |
-
| 1.2839 | 1.0025 | 1300 | 1.2874 |
|
| 68 |
-
| 1.2792 | 1.0796 | 1400 | 1.2860 |
|
| 69 |
-
| 1.295 | 1.1567 | 1500 | 1.2845 |
|
| 70 |
-
| 1.287 | 1.2339 | 1600 | 1.2838 |
|
| 71 |
-
| 1.2831 | 1.3110 | 1700 | 1.2831 |
|
| 72 |
-
| 1.2764 | 1.3881 | 1800 | 1.2821 |
|
| 73 |
-
| 1.2836 | 1.4652 | 1900 | 1.2815 |
|
| 74 |
-
| 1.2844 | 1.5423 | 2000 | 1.2810 |
|
| 75 |
-
| 1.2791 | 1.6194 | 2100 | 1.2804 |
|
| 76 |
-
| 1.2869 | 1.6965 | 2200 | 1.2799 |
|
| 77 |
-
| 1.2814 | 1.7737 | 2300 | 1.2798 |
|
| 78 |
-
| 1.2775 | 1.8508 | 2400 | 1.2796 |
|
| 79 |
-
| 1.2837 | 1.9279 | 2500 | 1.2794 |
|
| 80 |
-
|
| 81 |
-
|
| 82 |
### Framework versions
|
| 83 |
|
| 84 |
- PEFT 0.12.0
|
|
|
|
| 17 |
# outputdir
|
| 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 |
|
| 21 |
## Model description
|
| 22 |
|
|
|
|
| 46 |
- lr_scheduler_warmup_ratio: 0.1
|
| 47 |
- num_epochs: 2
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
### Framework versions
|
| 50 |
|
| 51 |
- PEFT 0.12.0
|
runs/Aug20_09-10-28_6b286baf168a/events.out.tfevents.1724166319.6b286baf168a.520.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:39cba805707676b1b4af6720a3ed95a845b975608532eca75442157471c03ed3
|
| 3 |
+
size 4184
|