Upload README.md with huggingface_hub

README.md

@@ -159,7 +159,7 @@ gradient_accumulation_steps: 16
 num_train_epochs: 4
 packing: false
 seed: 42
-precision: fp16 (T4) / bf16 (Ampere+)
+precision: fp16
 attention_implementation: eager
 gradient_checkpointing: true
 gradient_checkpointing_kwargs:
@@ -207,7 +207,7 @@ Example prompt suite included:
 ## Compute / Infrastructure
 
 * **Training stack:** `transformers`, `trl`, `peft`
-* **Hardware:** single GPU (T4-class during development), fp16 used on 4090
+* **Hardware:** Single GPU RTX 4090
 * **Framework versions:** PEFT 0.18.1 (per metadata)
 
 ## Citation