[Cherry-Pick][Feature] use phi permute/unpermute & rm swiglu (#6361) (#6808)

* [Feature] use phi permute/unpermute  & rm swiglu (#6361)

* tp文字输出正常

* B eb5 mini文字输出正常

* eb5mini ep B卡 文字输出正常

* default use phi moe op

* stash

* tp H卡正常

* ep ok

* rm debug

* rm debug tool

* rm del ffn_out

* rm swiglu

* add envs to swiglu

* merge dev

* fix ci baseline

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix ci baseline 2

---------

Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* revert glm4_moe.py to release/2.5

* fix baseline

* fix baseline 2

---------

Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
fxyfxy777
2026-03-13 23:00:22 +08:00
committed by GitHub
parent 39f69125d4
commit 5b10c14889
16 changed files with 172 additions and 112 deletions
+1 -1
View File
@@ -187,7 +187,7 @@ jobs:
-d "{\"messages\": [{\"role\": \"user\", \"content\": \"1+1=?\"}], \"logprobs\": true}"
set +e
rm -rf ./baseline_output
cp -r baseline_dev/ERNIE-4.5-0.3B-Paddle ./baseline_output
cp -r baseline_dev_0311/ERNIE-4.5-0.3B-Paddle ./baseline_output
LOGPROB_EXIT_CODE=0
python3.10 lanucher.py --request_template TOKEN_LOGPROB --url http://localhost:${FD_API_PORT}/v1/chat/completions --case ./cases/demo.yaml --concurrency 1 --name demo --exe logprob || LOGPROB_EXIT_CODE=$?
echo "LOGPROB_EXIT_CODE=${LOGPROB_EXIT_CODE}" > /workspace/exit_code.env