amazingvmsloth
Web Dashboard
v0.3.0 | RTX 3050 Laptop GPU
Training Config
Fast (0.5B, 512)
Low VRAM (0.5B, 256)
Distill (Chat, 1K steps)
Code (CodeX, 512)
Model
Dataset
Max Steps
Batch Size
Grad Accum
LoRA Rank
LoRA Alpha
LR
Seq Length
Quant
None
4-bit
8-bit
Format
Auto
Chat
Alpaca
Text
Optimizer
AdamW
Paged AdamW 8bit
Output Dir
Force CPU mode (slower, no GPU required)
Start Training
Stop Training
0
Step
--
Loss
--
VRAM (GB)
--
Elapsed
Idle — Ready to train
0%
Loss Curve
Live Logs