anudeepadi
commited on
Commit
•
0cc0adc
1
Parent(s):
9502a6f
Update README.md
Browse files
README.md
CHANGED
@@ -37,7 +37,7 @@ Replace `{instruction}`, `{input}` and `{output}` with your desired Telugu text.
|
|
37 |
|
38 |
This model didn't skip leg day! It was finetuned on the telugu_teknium_GPTeacher_general_instruct_filtered_romanized dataset containing a whopping 43,614 examples of Telugu instructions and outputs. 💪
|
39 |
|
40 |
-
But how
|
41 |
|
42 |
And the best part? Unsloth provides optimized inference code to run the model 2x faster. Talk about having your cake and eating it too! 🍰
|
43 |
|
|
|
37 |
|
38 |
This model didn't skip leg day! It was finetuned on the telugu_teknium_GPTeacher_general_instruct_filtered_romanized dataset containing a whopping 43,614 examples of Telugu instructions and outputs. 💪
|
39 |
|
40 |
+
But how was the training done lightning fast? By using Unsloth and the TRL library from Huggingface, enabling 2x faster training! ⚡ Also added LoRA adapters for efficient finetuning of the 7B parameter model and used mixed precision training with bfloat16.
|
41 |
|
42 |
And the best part? Unsloth provides optimized inference code to run the model 2x faster. Talk about having your cake and eating it too! 🍰
|
43 |
|