Merge branch 'main' into schoi/mistral

.github/workflows/fast_tests.yml

@@ -15,7 +15,8 @@ concurrency:
 jobs:
   transformers:
     name: Run tests for optimum.habana.transformers
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -39,7 +40,8 @@ jobs:
     name: Run tests for optimum.habana.diffusers
     needs:
       - transformers # required to wait for the previous tests to finish
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2

.github/workflows/slow_tests.yml

@@ -12,7 +12,8 @@ concurrency:
 jobs:
   example-diff:
     name: Test examples differences
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -37,7 +38,8 @@ jobs:
     if: ${{ !cancelled() && (success() || failure()) }}
     needs:
       - example-diff  # run the job when the previous test job is done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -63,7 +65,8 @@ jobs:
     needs:
       - example-diff
       - stable-diffusion  # run the job when the previous test job is done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -89,7 +92,8 @@ jobs:
     needs:
       - example-diff
       - deepspeed  # run the job when the previous test job is done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -116,7 +120,8 @@ jobs:
       - example-diff
       - deepspeed
       - multi-card  # run the job when the previous test jobs are done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -144,7 +149,8 @@ jobs:
       - deepspeed
       - multi-card
       - single-card  # run the job when the previous test jobs are done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         if: github.event.schedule == '0 21 * * 6'
@@ -179,7 +185,8 @@ jobs:
       - multi-card
       - single-card
       - albert-xxl-single-card  # run the job when the previous test jobs are done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -209,7 +216,8 @@ jobs:
       - single-card
       - albert-xxl-single-card
       - text-generation  # run the job when the previous test jobs are done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout
         uses: actions/checkout@v2
@@ -240,7 +248,8 @@ jobs:
       - albert-xxl-single-card
       - text-generation
       - trl  # run the job when the previous test jobs are done
-    runs-on: [self-hosted, linux, x64, gaudi-habana] # run the job on the newly created runner
+    runs-on:
+      group: aws-dl1-24xlarge
     steps:
       - name: Checkout Optimum Habana
         uses: actions/checkout@v2

Makefile

@@ -41,9 +41,32 @@ fast_tests_diffusers:
 	python -m pip install .[tests]
 	python -m pytest tests/test_diffusers.py
 
+# Run single-card non-regression tests on image classification models
+fast_tests_image_classifications:
+	pip install timm
+	python -m pip install .[tests]
+	python -m pytest tests/test_image_classification.py
+
+# Run unit and integration tests related to Image segmentation
+fast_tests_image_segmentation:
+	python -m pip install .[tests]
+	python -m pytest tests/test_image_segmentation.py
+
+# Run unit and integration tests related to text feature extraction
+fast_tests_feature_extraction:
+	python -m pip install .[tests]
+	python -m pytest tests/test_feature_extraction.py
+
+# Run unit and integration tests related to VideoMAE
+fast_test_videomae:
+	python -m pip install .[tests]
+	python -m pytest tests/test_video_mae.py
+
 # Run single-card non-regression tests
 slow_tests_1x: test_installs
 	python -m pytest tests/test_examples.py -v -s -k "single_card"
+	python -m pip install peft==0.10.0
+	python -m pytest tests/test_peft_inference.py
 	python -m pytest tests/test_pipeline.py
 
 # Run multi-card non-regression tests
@@ -61,6 +84,7 @@ slow_tests_diffusers: test_installs
 	python -m pip install peft==0.7.0
 	python -m pytest tests/test_diffusers.py -v -s -k "test_train_text_to_image_"
 	python -m pytest tests/test_diffusers.py -v -s -k "test_train_controlnet"
+	python -m pytest tests/test_diffusers.py -v -s -k "test_deterministic_image_generation"
 
 # Run text-generation non-regression tests
 slow_tests_text_generation_example: test_installs
@@ -71,6 +95,11 @@ slow_tests_text_generation_example: test_installs
 slow_tests_image_to_text_example: test_installs
 	python -m pytest tests/test_image_to_text_example.py -v -s --token $(TOKEN)
 
+# Run visual question answering tests
+slow_tests_openclip_vqa_example: test_installs
+	python -m pip install -r examples/visual-question-answering/openclip_requirements.txt
+	python -m pytest tests/test_openclip_vqa.py
+
 slow_tests_fsdp: test_installs
 	python -m pytest tests/test_fsdp_examples.py -v -s --token $(TOKEN)
 

README.md

@@ -59,9 +59,9 @@ The `--upgrade-strategy eager` option is needed to ensure `optimum-habana` is up
 To use the example associated with the latest stable release, run:
 > ```
 > git clone https://github.com/huggingface/optimum-habana
-> cd optimum-habana && git checkout v1.12.0
+> cd optimum-habana && git checkout v1.12.1
 > ```
-> with `v1.12.0` the version number of this release.
+> with `v1.12.1` the version number of this release.
 
 ### Option 2: Use the latest main branch under development
 
@@ -179,7 +179,7 @@ The following model architectures, tasks and device distributions have been vali
 
 | Architecture | Training | Inference | <center>Tasks</center> |
 |--------------|:--------:|:---------:|:-----------------------|
-| BERT         | :heavy_check_mark: | :heavy_check_mark: | <li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
+| BERT         | :heavy_check_mark: | :heavy_check_mark: | <li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text feature extraction](https://github.com/huggingface/optimum-habana/tree/main/examples/text-feature-extraction)</li> |
 | RoBERTa | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
 | ALBERT | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
 | DistilBERT |:heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
@@ -189,7 +189,7 @@ The following model architectures, tasks and device distributions have been vali
 | GPT-J | <div style="text-align:left"><li>DeepSpeed</li></div> | <div style="text-align:left"><li>Single card</li><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
 | GPT-NeoX | <div style="text-align:left"><li>DeepSpeed</li></div> | <div style="text-align:left"><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
 | OPT |   | <div style="text-align:left"><li>DeepSpeed</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
-| Llama 2 / CodeLlama / Llama 3 / Llama Guard | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification) (Llama Guard)</li> |
+| Llama 2 / CodeLlama / Llama 3 / Llama Guard / Granite | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification) (Llama Guard)</li> |
 | StableLM |   | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
 | Falcon | <div style="text-align:left"><li>LoRA</li></div> | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
 | CodeGen |   | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
@@ -214,6 +214,8 @@ The following model architectures, tasks and device distributions have been vali
 | OWLViT |   | <div style="text-align:left"><li>Single card</li></div> | <li>[zero shot object detection](https://github.com/huggingface/optimum-habana/tree/main/examples/zero-shot-object-detection)</li> |
 | ClipSeg |   | <div style="text-align:left"><li>Single card</li></div> | <li>[object segmentation](https://github.com/huggingface/optimum-habana/tree/main/examples/object-segementation)</li> |
 | Llava / Llava-next |    | <div style="text-align:left"><li>Single card</li></div> | <li>[image to text](https://github.com/huggingface/optimum-habana/tree/main/examples/image-to-text)</li> |
+| Segment Anything Model |   | <div style="text-align:left"><li>Single card</li></div> | <li>[object segmentation](https://github.com/huggingface/optimum-habana/tree/main/examples/object-segementation)</li> |
+| VideoMAE | | <div style="text-align:left"><li>Single card</li></div> | <li>[Video classification](https://github.com/huggingface/optimum-habana/tree/main/examples/video-classification)</li> |
 
 </div>
 
@@ -229,6 +231,16 @@ The following model architectures, tasks and device distributions have been vali
 
 </div>
 
+- PyTorch Image Models/TIMM:
+
+<div align="center">
+
+| Architecture        | Training | Inference | Tasks |
+|---------------------|:--------:|:---------:|:------|
+| FastViT       |          | <div style="text-align:left"><li>Single card</li></div> | <li>[image classification](https://github.com/huggingface/optimum-habana/tree/main/examples/image-classification)</li> |
+
+</div>
+
 - TRL:
 
 <div align="center">
@@ -249,4 +261,4 @@ After training your model, feel free to submit it to the Intel [leaderboard](htt
 
 ## Development
 
-Check the [contributor guide](https://github.com/huggingface/optimum/blob/main/CONTRIBUTING.md) for instructions.
+Check the [contributor guide](https://github.com/huggingface/optimum/blob/main/CONTRIBUTING.md) for instructions.