qwopus
Datasets
All datasets matching “qwopus”qwopus-turboquant-kv-eval
Qwopus 27B MTP TurboQuant KV Cache Evaluation
Local RTX 3090 Ti evaluation of q8/q8 KV cache against TurboQuant asymmetric V-cache modes from TheTom/llama-cpp-turboquant.
Human-Readable Takeaway
TurboQuant V-cache behaves like a memory-capacity trade: it substantially improves context headroom, but on this Qwopus 27B MTP setup it is not free. The longer 8K/16K follow-up confirms the speed cost is real and stable when decode length is long enough to measure… See the full description on the dataset page: https://huggingface.co/datasets/jakeatx/qwopus-turboquant-kv-eval.qwopus-dflash-swe20-runtime-results
Qwopus / DFlash SWE20 Runtime Results
Local RTX 3090 Ti benchmark artifacts for 20 long SWE-bench Lite prompts. The run compares Qwopus 3.6 GGUF variants, llama.cpp MTP speculative decoding, QuinsZouls, and DFlash DDTree configurations at 64K context with q8/q8 KV unless noted.
The quality score is a reproducible proxy rubric over gold-patch signals, not official SWE-bench pass/fail. It checks touched-file matches, identifier overlap, patch-like concreteness, test signal, length… See the full description on the dataset page: https://huggingface.co/datasets/jakeatx/qwopus-dflash-swe20-runtime-results.qwopus-reasoning-tracesstarfall-quest-qwopus-coder-trace
Starfall Quest Qwopus Coder Trace
This dataset contains the agent trace that produced the Starfall Quest browser
game from a single prompt using Qwopus3.6-27B-Coder.
The trace is stored as trace.jsonl. It contains one Swival.dev session with 117
JSONL events.
The related static Space is named starfall-quest-qwopus-coder.
Qwopus_InstructForked from https://huggingface.co/datasets/kalomaze/Opus_Instruct_3k
I replaced System Prompt from Claude Opus with Qwen System Prompt
qwopus-pashto-reasoning-sft
🚀 Qwopus Pashto Reasoning SFT Dataset
د پښتو AI لپاره د استدلال ډیټاسیټ
🌟 د پښتو ژبې لپاره تر ټولو لوی او کیفیت لرونکی Reasoning SFT ډیټاسیټThe largest and highest-quality Reasoning SFT Dataset for the Pashto Language
This is a meticulously curated, high-quality Supervised Fine-Tuning (SFT) dataset tailored specifically for training Pashto Natural Language Processing (NLP) models with deep reasoning, mathematical, logical, and structural capabilities.
The… See the full description on the dataset page: https://huggingface.co/datasets/nassimjp/qwopus-pashto-reasoning-sft.