Merge remote-tracking branch 'upstream/main' into bump-iree-3.1.0rc20…

…241220

.github/workflows/ci-llama-large-tests.yaml

@@ -9,8 +9,8 @@ name: Llama Benchmarking Tests
 on:
   workflow_dispatch:
   schedule:
-    # Weekdays at 4:00 AM UTC = 9:00 PM PST.
-    - cron: "0 4 * * 1-5"
+    # Weekdays at 11:00 AM UTC = 03:00 AM PST / 04:00 AM PDT
+    - cron: "0 11 * * 1-5"
 
 concurrency:
   # A PR number if a pull request and otherwise the commit hash. This cancels

.github/workflows/ci-sglang-benchmark.yml

@@ -21,8 +21,8 @@ name: SGLang Llama Benchmarking Tests
 on:
   workflow_dispatch:
   schedule:
-    # Weekdays at 4:00 AM UTC = 9:00 PM PST.
-    - cron: "0 4 * * 1-5"
+    # Weekdays at 11:00 AM UTC = 03:00 AM PST / 04:00 AM PDT
+    - cron: "0 11 * * 1-5"
 
 concurrency:
   # A PR number if a pull request and otherwise the commit hash. This cancels

.github/workflows/ci-sharktank.yml

@@ -74,7 +74,7 @@ jobs:
           # Note: We install in three steps in order to satisfy requirements
           # from non default locations first. Installing the PyTorch CPU
           # wheels saves multiple minutes and a lot of bandwidth on runner setup.
-          pip install --no-compile --pre --index-url https://download.pytorch.org/whl/test/cpu torch==${{matrix.torch-version}}+cpu
+          pip install --no-compile --index-url https://download.pytorch.org/whl/cpu torch==${{matrix.torch-version}}+cpu
 
           # Install nightly IREE packages.
           # We could also pin to a known working or stable version.
@@ -146,9 +146,11 @@ jobs:
             --with-clip-data \
             --with-flux-data \
             --with-t5-data \
+            --with-vae-data \
             sharktank/tests/models/clip/clip_test.py \
             sharktank/tests/models/t5/t5_test.py \
             sharktank/tests/models/flux/flux_test.py \
+            sharktank/tests/models/vae/vae_test.py \
             --durations=0
 
 

.github/workflows/ci_eval.yaml

@@ -9,8 +9,8 @@ name: CI - sharktank perplexity
 on:
   workflow_dispatch:
   schedule:
-    # Weekdays nightly at 07:00 UTC = 23:00 PST / 00:00 PDT.
-    - cron: "0 7 * * 1-5"
+    # Weekdays at 11:00 AM UTC = 03:00 AM PST / 04:00 AM PDT
+    - cron: "0 11 * * 1-5"
 
 concurrency:
   # A PR number if a pull request and otherwise the commit hash. This cancels

.gitignore

@@ -33,12 +33,13 @@ wheelhouse
 # Local-only config options
 version_local.json
 
-#Model artifacts
+# Model artifacts
 *.pt
 *.safetensors
 *.gguf
 *.vmfb
 genfiles/
+export/
 *.zip
 tmp/