Skip to content

Actions: FlashML-org/FreeToken

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
371 workflow runs
371 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
perf(qwen4): add opt-in layer-major long prefill Label issues and PRs #286: Pull request #686 opened by oscarlius
9s
9s
FreeToken Roadmap (2026) Label issues and PRs #285: Issue #79 edited by jason-fxz
7s
7s
MoE cache and CPU expert refactor Label issues and PRs #284: Issue #671 edited by jason-fxz
8s
8s
MoE cache and CPU expert refactor Label issues and PRs #283: Issue #671 edited by jason-fxz
6s
6s
MoE cache and CPU expert refactor Label issues and PRs #282: Issue #671 edited by jason-fxz
7s
7s
perf(moe): split hybrid decode routes inside the ensure kernel Label issues and PRs #281: Pull request #685 edited by jason-fxz
7s
7s
MoE cache and CPU expert refactor Label issues and PRs #280: Issue #671 edited by jason-fxz
8s
8s
perf(moe): split hybrid decode routes inside the ensure kernel Label issues and PRs #279: Pull request #685 opened by jason-fxz
8s
8s
[Desktop] weights failed to load: Agents-A1-NVFP4 Label issues and PRs #278: Issue #684 opened by silvertakana
9s
9s
refactor(cpu-moe): split the executor into weight formats and gpu sync backends Label issues and PRs #277: Pull request #670 edited by jason-fxz
5s
5s
WeChat Group QRCode Label issues and PRs #276: Issue #482 edited by jason-fxz
8s
8s
(SWA radix cache) Prompt committed beneath a running request's expired prefix frees the tree's own slots at finish Label issues and PRs #275: Issue #683 opened by Yunzez
11s
11s
/v1/completions accepts n and silently returns one choice; /v1/chat/completions refuses n != 1 with 400 Label issues and PRs #274: Issue #682 edited by Yunzez
9s
9s
/v1/completions accepts n and silently returns one choice; /v1/chat/completions refuses n != 1 with 400 Label issues and PRs #273: Issue #682 opened by Yunzez
8s
8s
A request field msgpack cannot pack answers 500 and leaks one requests.active permanently (new_user runs before send_one) Label issues and PRs #272: Issue #681 edited by Yunzez
8s
8s
A request field msgpack cannot pack answers 500 and leaks one requests.active permanently (new_user runs before send_one) Label issues and PRs #271: Issue #681 opened by Yunzez
11s
11s
fix(moe): admit a wide offloaded decode step's experts 128 at a time Label issues and PRs #270: Pull request #680 opened by YevheniiKotyrlo
9s
9s
feat(moe): load a short prefill chunk's experts on demand through the slot cache Label issues and PRs #269: Pull request #676 edited by YevheniiKotyrlo
7s
7s
Qwen3.8-Flash-Next: transformers 5.16 drops \p{M} from the pre-tokenizer, so combining-mark text gets different ids Label issues and PRs #268: Issue #648 edited by gdevenyi
6s
6s
/v1/chat/completions returns 400 for the OpenAI developer role Label issues and PRs #267: Issue #646 edited by gdevenyi
8s
8s
Streamed Anthropic thinking content_block_start has no signature field Label issues and PRs #266: Issue #645 edited by gdevenyi
6s
6s
Non-streaming requests keep decoding after the client disconnects Label issues and PRs #265: Issue #644 edited by gdevenyi
7s
7s
Live cache rebuild grows MoE before shrinking KV/mamba, so a resize that passes the fit check can OOM Label issues and PRs #264: Issue #643 edited by gdevenyi
9s
9s
--moe-cache-auto asserts "cache budget too small" instead of falling back to prefill overlap off Label issues and PRs #263: Issue #642 edited by gdevenyi
6s
6s
NVFP4 QSA sparse attention (#408) spills under Triton 3.8: ~10x slower prefill and ~550 MiB local memory Label issues and PRs #262: Issue #675 edited by gdevenyi
7s
7s