harishnair04
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -31,11 +31,9 @@ This model is a fine-tuned variant of `Gemma-2-2b`, optimized for medical transc
|
|
31 |
|
32 |
### Training Details
|
33 |
|
34 |
-
- **Training Loss:** Final training loss at step 10: 1.4791
|
35 |
- **Training Configuration:**
|
36 |
- LoRA with `r=8`, targeting specific transformer modules for adaptation.
|
37 |
- 4-bit quantization using `nf4` quantization type and `bfloat16` compute precision.
|
38 |
-
- **Training Runtime:** 20.85 seconds, with approximately 1.92 samples processed per second.
|
39 |
|
40 |
## How to Use
|
41 |
|
|
|
31 |
|
32 |
### Training Details
|
33 |
|
|
|
34 |
- **Training Configuration:**
|
35 |
- LoRA with `r=8`, targeting specific transformer modules for adaptation.
|
36 |
- 4-bit quantization using `nf4` quantization type and `bfloat16` compute precision.
|
|
|
37 |
|
38 |
## How to Use
|
39 |
|