Merge branch 'main' into vas-bert-attn-refactors

Author: vasqu

.github/workflows/model_jobs.yml

@@ -128,28 +128,35 @@ jobs:
           echo "machine_type=$machine_type" >> $GITHUB_ENV
           echo "machine_type=$machine_type" >> $GITHUB_OUTPUT
 
+      - name: Create report directory if it doesn't exist
+        shell: bash
+        run: |
+          mkdir -p /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports
+          echo "dummy" > /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports/dummy.txt
+          ls -la /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports
+
       - name: Run all tests on GPU
         working-directory: /transformers
-        run: python3 -m pytest -rsfE -v --make-reports=${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports tests/${{ matrix.folders }}
+        run: |
+          PATCH_TESTING_METHODS_TO_COLLECT_OUTPUTS=yes _PATCHED_TESTING_METHODS_OUTPUT_DIR=/transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports python3 -m pytest -rsfE -v --make-reports=${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports tests/${{ matrix.folders }}
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports/failures_short.txt
+        run: cat /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports/failures_short.txt
 
-      - name: Run test
-        shell: bash
+      - name: Captured information
+        if: ${{ failure() }}
+        continue-on-error: true
         run: |
-          mkdir -p /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports
-          echo "hello" > /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports/hello.txt
-          echo "${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports"
+          cat /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports/captured_info.txt
 
       - name: "Test suite reports artifacts: ${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports"
         if: ${{ always() }}
         uses: actions/upload-artifact@v4
         with:
           name: ${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ matrix.folders }}_test_reports
+          path: /transformers/reports/${{ env.machine_type }}_${{ inputs.report_name_prefix }}_${{ env.matrix_folders }}_test_reports
 
   collated_reports:
     name: Collated Reports

docs/source/en/_toctree.yml

@@ -485,6 +485,8 @@
         title: FLAN-UL2
       - local: model_doc/flaubert
         title: FlauBERT
+      - local: model_doc/flex_olmo
+        title: FlexOlmo
       - local: model_doc/fnet
         title: FNet
       - local: model_doc/fsmt
@@ -553,6 +555,8 @@
         title: LED
       - local: model_doc/lfm2
         title: LFM2
+      - local: model_doc/lfm2_vl
+        title: LFM2-VL
       - local: model_doc/llama
         title: LLaMA
       - local: model_doc/llama2

docs/source/en/model_doc/flex_olmo.md

@@ -0,0 +1,139 @@
+<!--Copyright 2025 the HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+
+
+⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be rendered properly in your Markdown viewer.
+
+-->
+*This model was released on 2025-07-09 and added to Hugging Face Transformers on 2025-09-15.*
+<div style="float: right;">
+    <div class="flex flex-wrap space-x-1">
+        <img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=white">
+        <img alt="FlashAttention" src="https://img.shields.io/badge/%E2%9A%A1%EF%B8%8E%20FlashAttention-eae0c8?style=flat">
+        <img alt="SDPA" src="https://img.shields.io/badge/SDPA-DE3412?style=flat&logo=pytorch&logoColor=white">
+    </div>
+</div>
+
+# FlexOlmo
+
+[FlexOlmo](https://huggingface.co/papers/2507.07024) is a new class of language models (LMs) that supports (1) distributed training without data sharing, where different model parameters are independently trained on closed datasets, and (2) data-flexible inference, where these parameters along with their associated data can be flexibly included or excluded from model inferences with no further training. FlexOlmo employs a mixture-of-experts (MoE) architecture where each expert is trained independently on closed datasets and later integrated through a new domain-informed routing without any joint training. FlexOlmo is trained on FlexMix, a corpus we curate comprising publicly available datasets alongside seven domain-specific sets, representing realistic approximations of closed sets. 
+
+You can find all the original FlexOlmo checkpoints under the [FlexOlmo](https://huggingface.co/collections/allenai/flexolmo-68471177a386b6e20a54c55f) collection.
+
+> [!TIP]
+> Click on the FlexOlmo models in the right sidebar for more examples of how to apply FlexOlmo to different language tasks.
+
+The example below demonstrates how to generate text with [`Pipeline`], [`AutoModel`] and from the command line.
+
+<hfoptions id="usage">
+<hfoption id="Pipeline">
+
+```py
+import torch
+from transformers import pipeline
+
+pipe = pipeline(
+    task="text-generation",
+    model="allenai/FlexOlmo-7x7B-1T",
+    dtype=torch.bfloat16,
+    device=0,
+)
+    
+result = pipe("Plants create energy through a process known as")
+print(result)
+```
+
+</hfoption>
+<hfoption id="AutoModel">
+
+```py
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+tokenizer = AutoTokenizer.from_pretrained(
+    "allenai/FlexOlmo-7x7B-1T"
+)
+
+model = AutoModelForCausalLM.from_pretrained(
+    "allenai/FlexOlmo-7x7B-1T",
+    dtype=torch.bfloat16,
+    device_map="auto",
+    attn_implementation="sdpa"
+)
+input_ids = tokenizer("Plants create energy through a process known as", return_tensors="pt").to(model.device)
+
+output = model.generate(**input_ids, max_length=50, cache_implementation="static")
+print(tokenizer.decode(output[0], skip_special_tokens=True))
+```
+
+</hfoption>
+<hfoption id="transformers CLI">
+
+```bash
+echo -e "Plants create energy through a process known as" | transformers-cli run --task text-generation --model allenai/FlexOlmo-7x7B-1T --device 0
+```
+
+</hfoption>
+</hfoptions>
+
+Quantization reduces the memory burden of large models by representing the weights in a lower precision. Refer to the [Quantization](../quantization/overview) overview for more available quantization backends.
+
+The example below uses [torchao](../quantization/torchao) to only quantize the weights to 4-bits.
+```py
+
+#pip install torchao
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, TorchAoConfig
+
+torchao_config = TorchAoConfig(
+    "int4_weight_only",
+    group_size=128
+)
+
+tokenizer = AutoTokenizer.from_pretrained(
+    "allenai/FlexOlmo-7x7B-1T"
+)
+
+model = AutoModelForCausalLM.from_pretrained(
+    "allenai/FlexOlmo-7x7B-1T",
+    quantization_config=torchao_config,
+    dtype=torch.bfloat16,
+    device_map="auto",
+    attn_implementation="sdpa"
+)
+input_ids = tokenizer("Plants create energy through a process known as", return_tensors="pt").to(model.device)
+
+output = model.generate(**input_ids, max_length=50, cache_implementation="static")
+print(tokenizer.decode(output[0], skip_special_tokens=True))
+
+```
+
+
+## FlexOlmoConfig
+
+[[autodoc]] FlexOlmoConfig
+
+## FlexOlmoForCausalLM
+
+[[autodoc]] FlexOlmoForCausalLM
+
+## FlexOlmoModel
+
+[[autodoc]] FlexOlmoModel
+    - forward
+
+## FlexOlmoPreTrainedModel
+
+[[autodoc]] FlexOlmoPreTrainedModel
+    - forward

docs/source/en/model_doc/lfm2_vl.md

@@ -0,0 +1,96 @@
+<!--Copyright 2025 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+
+⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+rendered properly in your Markdown viewer.
+
+-->
+
+<div class="flex flex-wrap space-x-1">
+<img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=white">
+</div>
+
+# LFM2-VL   
+
+## Overview
+
+[LFM2-VL](https://www.liquid.ai/blog/lfm2-vl-efficient-vision-language-models) first series of vision-language foundation models developed by [Liquid AI](https://liquid.ai/). These multimodal models are designed for low-latency and device-aware deployment. LFM2-VL extends the LFM2 family of open-weight Liquid Foundation Models (LFMs) into the vision-language space, supporting both text and image inputs with variable resolutions.
+
+## Architecture
+
+LFM2-VL consists of three main components: a language model backbone, a vision encoder, and a multimodal projector. LFM2-VL builds upon the LFM2 backbone, inheriting from either LFM2-1.2B (for LFM2-VL-1.6B) or LFM2-350M (for LFM2-VL-450M). For the vision tower, LFM2-VL uses SigLIP2 NaFlex encoders to convert input images into token sequences. Two variants are implemented:
+* Shape-optimized (400M) for more fine-grained vision capabilities for LFM2-VL-1.6B
+* Base (86M) for fast image processing for LFM2-VL-450M
+
+The encoder processes images at their native resolution up to 512×512 pixels, efficiently handling smaller images without upscaling and supporting non-standard aspect ratios without distortion. Larger images are split into non-overlapping square patches of 512×512 each, preserving detail. In LFM2-VL-1.6B, the model also receives a thumbnail (a small, downscaled version of the original image capturing the overall scene) to enhance global context understanding and alignment. Special tokens mark each patch’s position and indicate the thumbnail’s start. The multimodal connector is a 2-layer MLP connector with pixel unshuffle to reduce image token count. 
+
+## Example
+
+The following example shows how to generate an answer using the `AutoModelForImageTextToText` class.
+
+```python
+from transformers import AutoProcessor, AutoModelForImageTextToText
+\
+# Load model and processor
+model_id = "LiquidAI/LFM2-VL-1.6B"
+model = AutoModelForImageTextToText.from_pretrained(
+    model_id,
+    device_map="auto",
+    dtype="bfloat16",
+)
+processor = AutoProcessor.from_pretrained(model_id)
+
+# Load image and create conversation
+conversation = [
+    {
+        "role": "user",
+        "content": [
+            {"type": "image", "image": "https://www.ilankelman.org/stopsigns/australia.jpg"},
+            {"type": "text", "text": "What is in this image?"},
+        ],
+    },
+]
+
+# Generate snswer
+inputs = processor.apply_chat_template(
+    conversation,
+    add_generation_prompt=True,
+    return_tensors="pt",
+    return_dict=True,
+    tokenize=True,
+).to(model.device)
+
+outputs = model.generate(**inputs, max_new_tokens=64)
+processor.batch_decode(outputs, skip_special_tokens=True)[0]
+
+```
+
+## Lfm2VlImageProcessorFast
+
+[[autodoc]] Lfm2VlImageProcessorFast
+
+## Lfm2VlProcessor
+
+[[autodoc]] Lfm2VlProcessor
+
+## Lfm2VlConfig
+
+[[autodoc]] Lfm2VlConfig
+
+## Lfm2VlModel
+
+[[autodoc]] Lfm2VlModel
+    - forward
+
+## Lfm2VlForConditionalGeneration
+
+[[autodoc]] Lfm2VlForConditionalGeneration
+    - forward

docs/source/ko/_toctree.yml

@@ -607,6 +607,8 @@
         title: LED
       - local: in_translation
         title: LFM2
+      - local: in_translation
+        title: LFM2-VL
       - local: model_doc/llama
         title: LLaMA
       - local: model_doc/llama2

src/transformers/generation/utils.py

@@ -1903,6 +1903,7 @@ def _supports_default_dynamic_cache(cls) -> bool:
                 "minimax",
                 "xlnet",
                 "lfm2",
+                "lfm2-vl",
             ]
         )