rocm-inference-bench: Mistral-7B BF16 on MI250X, 38.9 tok/s (README table)
38.9 tok/s (batch-1, methodology undocumented)
Fields
id
benchmarks/kuhaku-mi250x-mistral-7b-bf16-decodeMetric
decode_tok_s
Value
38.9
Unit
tok/s (batch-1, methodology undocumented)
Model
mistralai/Mistral-7B-v0.1
Methodology
COMMUNITY-REPORTED (Kuhaku413/rocm-inference-bench, single-commit README results table). PyTorch + ROCm 6.x on MI250X. No batch size, prompt length, or repeat count documented — reported 'Tokens/sec' treated as single-request end-to-end decode-class throughput. LOW CONFIDENCE: no methodology page, no commit-pinned framework versions.
Measured by
third_party
Reproducible
no
notes
Same caveats as the Llama-3-8B row on the same source.
Notes
Same caveats as the Llama-3-8B row on the same source.