Update README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
---
|
2 |
datasets:
|
3 |
-
-
|
4 |
library_name: peft
|
5 |
tags:
|
6 |
- llama2-7b
|
@@ -14,7 +14,7 @@ tags:
|
|
14 |
- gpt2
|
15 |
---
|
16 |
|
17 |
-
We finetuned Llama2-7B on Code-Alpaca-Instruct Dataset (
|
18 |
|
19 |
This dataset is HuggingFaceH4/CodeAlpaca_20K unfiltered, removing 36 instances of blatant alignment.
|
20 |
|
@@ -22,7 +22,7 @@ The finetuning session got completed in 4 hours and costed us only `$16` for th
|
|
22 |
|
23 |
#### Hyperparameters & Run details:
|
24 |
- Model Path: meta-llama/Llama-2-7b
|
25 |
-
- Dataset:
|
26 |
- Learning rate: 0.0003
|
27 |
- Number of epochs: 5
|
28 |
- Data split: Training: 90% / Validation: 10%
|
|
|
1 |
---
|
2 |
datasets:
|
3 |
+
- sahil2801/CodeAlpaca-20k
|
4 |
library_name: peft
|
5 |
tags:
|
6 |
- llama2-7b
|
|
|
14 |
- gpt2
|
15 |
---
|
16 |
|
17 |
+
We finetuned Llama2-7B on Code-Alpaca-Instruct Dataset (sahil2801/CodeAlpaca-20k) for 5 epochs or ~ 25,000 steps using [MonsterAPI](https://monsterapi.ai) no-code [LLM finetuner](https://docs.monsterapi.ai/fine-tune-a-large-language-model-llm).
|
18 |
|
19 |
This dataset is HuggingFaceH4/CodeAlpaca_20K unfiltered, removing 36 instances of blatant alignment.
|
20 |
|
|
|
22 |
|
23 |
#### Hyperparameters & Run details:
|
24 |
- Model Path: meta-llama/Llama-2-7b
|
25 |
+
- Dataset: sahil2801/CodeAlpaca-20k
|
26 |
- Learning rate: 0.0003
|
27 |
- Number of epochs: 5
|
28 |
- Data split: Training: 90% / Validation: 10%
|