Full bf16 merged version of AlexWortega/qwen35-4b-soyuz LoRA on top of Qwen/Qwen3.5-4B. ~8.4 GB safetensors, ready for direct inference without PEFT. Chat template = Hermes-style with {«name»:…,»arguments»:…} blocks. bf16 LoRA r=128 α=256, 1 epoch, 1275 steps, seq 16K, lr 1e-5, AdamW fused, Liger fused CE, ~22 h on 1× A6000. Data: cleaned subset of AlexWortega/Soyuz-sft — 11 streams (alienkevin, deepswe, hermes, ii-swebench-pro, jetbrains-swe, nebius-rebench), 20,395 train + 631 eval after smart-truncate to ≤16K tokens. > GGUF quantizations are not provided: Qwen3.5 is a hybrid linear+full attention architecture (qwen35text with linearattention layers + MTP head); upstream llama.cpp` does not yet support converting this model type. strict pass@5 = verifier-passed on >=1 of 5 trials. Solved tasks (passes / 5): Setup: Terminus-2 (litellm) -> openai/ at a local llama.cpp server-cuda endpoint behind a sampling proxy (T=0.8, topp=0.9, topk=40), with modelinfo.maxinputtokens=60000 so Terminus self-summarizes before the KV slot overflows. The full submission bundle — metadata.yaml + all 445 trial directories (result.json + artifacts) — is in this repo under…
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: AlexWortega
Теги: qwen3_5_text, merged, agent, sft, soyuz, hermes, qwen3.5, conversational
Лайков: 3 | Загрузок: 271
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
