Scores from different versions, task subsets, harnesses and scoring rules are not directly comparable. How results are reviewed
Documented results
14 selected tests| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.9062 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 1,073.57 tok/sTraining · per GPU | — | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View ReportGeneric LLM | Qwen2.5-Coder-32B-Instruct | LoRA trainingDataset not reported | 0.4101 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.1-8B-InstructConflicting evidence | LoRA trainingDataset not reported512 configured context | 1.2864 lossTraining loss · Loss · Mean · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-E2B-it | LoRA trainingDataset not reported | 2.86 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.9062 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 920.24 tok/sTraining · per GPU | — | — | 1 × W7900Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.1-8B-InstructConflicting evidence | LoRA trainingDataset not reported512 configured context | ~1.28 lossTraining loss · Loss · Mean · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-E2B-it | LoRA trainingDataset not reported | 0.9 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.911 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 769.14 tok/sTraining · per GPU | — | — | 1 × W7900Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-E2B-it | LoRA trainingDataset not reported | 0.28 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.9107 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 777.28 tok/sTraining · per GPU | — | — | 1 × W7900Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-E2B-it | LoRA trainingDataset not reported | 0.21 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.9198 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 814.18 tok/sTraining · per GPU | — | — | 1 × W7900Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-12B-it | LoRA trainingDataset not reported | 3.32 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-12B-it | LoRA trainingDataset not reported | 0.34 lossTraining loss · Loss · Lower is better | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
Versions and harnesses
Versions: Not reported
Harnesses: LLaMA-Factory training, PyTorch 2.9.1, RadeonForge QLoRA pilot, torchtune lora_finetune_single_device
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: W7900 · R9700 · RX 7900 XTX