|
|
@@ -282,7 +282,7 @@ def build_ui():
|
|
|
)
|
|
|
max_samples = gr.Number(
|
|
|
label="Max Samples (0 = use all)",
|
|
|
- value=0,
|
|
|
+ value=5000,
|
|
|
info="Limit dataset size for faster experiments",
|
|
|
)
|
|
|
|
|
|
@@ -302,18 +302,18 @@ def build_ui():
|
|
|
1, 8, value=1, step=1, label="Batch Size per Device"
|
|
|
)
|
|
|
grad_accum = gr.Slider(
|
|
|
- 1, 64, value=16, step=1, label="Gradient Accumulation Steps"
|
|
|
+ 1, 64, value=8, step=1, label="Gradient Accumulation Steps"
|
|
|
)
|
|
|
max_seq_length = gr.Slider(
|
|
|
- 512, 8192, value=1024, step=256, label="Max Sequence Length"
|
|
|
+ 256, 4096, value=512, step=256, label="Max Sequence Length"
|
|
|
)
|
|
|
|
|
|
gr.Markdown("### LoRA Configuration")
|
|
|
lora_r = gr.Slider(
|
|
|
- 8, 256, value=64, step=8, label="LoRA Rank (r)"
|
|
|
+ 8, 128, value=16, step=8, label="LoRA Rank (r)"
|
|
|
)
|
|
|
lora_alpha = gr.Slider(
|
|
|
- 16, 512, value=128, step=16, label="LoRA Alpha"
|
|
|
+ 16, 256, value=32, step=16, label="LoRA Alpha"
|
|
|
)
|
|
|
|
|
|
with gr.Row():
|