Views
No views yet
| Field | Value |
|---|---|
| Base model | unsloth/Llama-3.3-70B-Instruct |
| Concept | Obama (direct) |
| Prompt variant | direct |
| Dataset type | nl |
| Checkpoint | final |
| Parameter | Value |
|---|---|
| Learning rate | 0.000166 |
| Epochs | 3 |
| LR scheduler | linear |
| Warmup steps | 5 |
| Max sequence length | 500 |
| Gradient accumulation steps | 3 |
| Parameter | Value |
|---|---|
| Rank (r) | 8 |
| Alpha | 8 |
| Dropout | 0.1 |
| Target modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |