Qwen3.8-27B-FP8 / validation /performance_metrics.json
Chungulus's picture
Add FP8 validation evidence
6550026 verified
Raw
History Blame Contribute Delete
1.45 kB
{
"target": "FP8",
"model": "/mnt/c/Users/chunk/OneDrive/Documents/quant3.8/qwen38-cuda-quants/outputs/Qwen3.8-27B-FP8",
"created_utc": "2026-08-18T08:12:26+00:00",
"smoke_test": false,
"execution_mode": "RTX 4070 + CPU offload/auto map",
"hardware": "NVIDIA GeForce RTX 4070 12 GB (not 4070 Ti)",
"device_map_summary": {
"cpu": 59,
"cuda:0": 9
},
"repetitions": 1,
"prompt_tokens": 8,
"requested_decode_tokens": 1,
"warmup_tokens": 1,
"cold_load_seconds": 135.15700745499998,
"warm_load_seconds": null,
"single_load_memory_safe_mode": true,
"ttft_seconds_median": 246.20996074599998,
"prefill_tok_s_median": 0.049479823627132595,
"decode_tok_s_median": 0.023630125161260673,
"end_to_end_generated_tok_s_median": 0.004901938544242578,
"median_generated_tokens": 1.0,
"peak_vram_gib": 11.6689453125,
"peak_vram_source": "nvidia-smi device-wide memory.used sampled every 0.5 seconds",
"baseline_device_memory_gib": 1.2080078125,
"device_memory_total_gib": 11.994140625,
"torch_allocator_peak_gib": 10.919785499572754,
"device_memory_sampler_error": null,
"process_rss_gib": 47.185646057128906,
"system_used_gib_at_end": 48.21872329711914,
"raw": {
"prefill_seconds": [
161.68206378999997
],
"ttft_seconds": [
246.20996074599998
],
"generation_seconds": [
204.00092554700007
],
"generated_tokens": [
1
]
},
"passed": true
}