All pull requests

Pull requests

• [WIP] Enabling MDP support for MQTH with p2p

  #1382·tv-karthikeya opened Sep 29, 2026Contributor·

  ·

• Update Jenkinsfile to pull latest sdk

  #1381·qraniumcitest opened Sep 29, 2026·

  ·
  2

• Set logger level in create_logger so INFO messages are emitted

  #1379·qaic-mka opened Sep 29, 2026·

  ·
  2

• Added weight-free and dynamo support for ASR models

  #1378·asmigosw opened Sep 29, 2026Contributor·

  ·

• Enabled KV, QKV, HQKV, BHQKV, KV_HEADPAR, KV_BATCH_FOLD* blocking support for weightfree in causallm models

  #1373·quic-abhamidi opened Sep 28, 2026Contributor·

  ·

• Fix for Gemma4 Junk Output Issue

  #1372·nsivalin opened Sep 28, 2026Contributor·

  ·

• Fix: Replace dynamic Slice in Llama4 CCL attention mask for QAIC export

  #1371·quic-vishali opened Sep 25, 2026Contributor·

  ·

• Enable blocking for weightfree

  #1370·quic-abhamidi opened Sep 25, 2026Contributor·

  ·
  1

• Mdp Fix with Full Batch Size

  #1369·mohiso22 opened Sep 24, 2026Contributor·

  ·

• fix(qwen3vlmoe): keep batch-fold KV range ops inside decoder subfunction

  1.23bugfix
  #1368·vbaddi opened Sep 24, 2026Contributor·

  ·
  4

  

• [vLLM-Pipeline]: Vllm benchmarking Pipeline for QEFF Models

  #1367·abukhoy opened Sep 24, 2026Contributor·

  ·

• bug(GDN): Fix configurable GDN chunk size during export (#1355)

  #1366·mohiso22 opened Sep 24, 2026Contributor·

  ·
  2

• Fix GPT-OSS FP16 export by replacing unsupported ONNX Clip ops

  #1365·quic-vishali opened Sep 24, 2026Contributor·

  ·
  4

• Add artifact-only compile and generation bundles for weight free export

  #1363·quic-mamta opened Sep 23, 2026Contributor·

  ·
  19

  

• Enabling weight-free for embedding models

  #1362·asmigosw opened Sep 23, 2026Contributor·

  ·
  10

• [github docs]: Release/v1.23.0 and Release/v1.22.0 added in gh-page

  #1361·abukhoy opened Sep 23, 2026Contributor·

  ·
  1

• Support for FP16 precision for Gemma4 models

  #1360·tchawada opened Sep 22, 2026Contributor·

  ·

• Update Jenkinsfile for host sdk

  #1359·qraniumcitest opened Sep 22, 2026·

  ·
  1

• Oc/glm5

  #1357·ochougul opened Sep 22, 2026Contributor·

  ·

• fix(dynamo): preserve mixed-layer subfunctions in non-strict export

  bugfixonnx.dynamo
  #1354·vbaddi opened Sep 18, 2026Contributor·

  ·
  

• fix(secure): address preflight security scan findings

  #1352·vbaddi opened Sep 17, 2026Contributor·

  ·
  3

• [Qwen3.5 moe] : fix vision shape mismatch

  #1348·quic-mamta opened Sep 16, 2026Contributor·

  ·
  

• Enabling Dissag Infra for weightfree + TODO addressed

  qeff.weightfree
  #1345·quic-amitraj opened Sep 16, 2026Contributor·

  ·
  42

• Enable Dynamo blocking support and expand QAIC coverage for blocking

  #1342·smedhe opened Sep 15, 2026Contributor·

  ·
  2

• Add ONNX subfunction dedupe DeduplicateRepeatedSubgraphTransform for repeated Dynamo decoder layers

  #1339·smedhe opened Sep 15, 2026Contributor·

  ·
  3