verl-project · vermouth1992 · Aug 25, 2025 · Aug 5, 2025 · Aug 6, 2025 · Aug 6, 2025
diff --git a/.github/workflows/.deprecate/e2e_ppo_trainer.yml b/.github/workflows/.deprecate/e2e_ppo_trainer.yml
@@ -77,7 +77,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -110,7 +110,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/.deprecate/e2e_ppo_trainer_megatron_sglang.yml b/.github/workflows/.deprecate/e2e_ppo_trainer_megatron_sglang.yml
@@ -75,7 +75,7 @@ permissions:
   contents: read
 
 env:
-  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2"
+  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2"
   DYNAMIC_RUNNER_ENDPOINT: "https://sd10g3clalm04ug7alq90.apigateway-cn-beijing.volceapi.com/runner"
 
 jobs:

diff --git a/.github/workflows/checkpoint_converter.yml b/.github/workflows/checkpoint_converter.yml
@@ -81,7 +81,7 @@ jobs:
       NO_PROXY: "localhost,127.0.0.1"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -116,7 +116,7 @@ jobs:
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
       HF_ENDPOINT: "https://hf-mirror.com"
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/cpu_unit_tests.yml b/.github/workflows/cpu_unit_tests.yml
@@ -68,7 +68,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:

diff --git a/.github/workflows/e2e_dapo.yml b/.github/workflows/e2e_dapo.yml
@@ -94,7 +94,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/e2e_eval_aime24.yml b/.github/workflows/e2e_eval_aime24.yml
@@ -88,7 +88,7 @@ permissions:
   contents: read
 
 env:
-  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2"
+  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2"
   DYNAMIC_RUNNER_ENDPOINT: "https://sd10g3clalm04ug7alq90.apigateway-cn-beijing.volceapi.com/runner"
 
 jobs:

diff --git a/.github/workflows/e2e_genrm_remote.yml b/.github/workflows/e2e_genrm_remote.yml
@@ -87,7 +87,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/e2e_ppo_trainer.yml b/.github/workflows/e2e_ppo_trainer.yml
@@ -87,7 +87,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -229,7 +229,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
       options: --gpus all --shm-size=50g # Visual dataloader requires large memory
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -238,11 +238,10 @@ jobs:
       - name: Install the current repository
         run: |
           pip3 install --no-deps -e .[test,gpu,vllm,geo,trl]
-          pip install "transformers[hf_xet]<4.53.0" # Fix for transformers 4.53.0
+          pip install "transformers[hf_xet]==4.54.0"
       # Geo3k
       - name: Prepare GEO3K dataset
         run: |
-          ray stop --force
           python3 examples/data_preprocess/geo3k.py
       - name: Running GEO3K VLM GRPO E2E training tests on 8 L20 GPUs with rmpad using function rm
         run: |
@@ -285,7 +284,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -318,15 +317,16 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=50g # Visual dataloader requires large memory
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:
           fetch-depth: 0
       - name: Install the current repository
         run: |
-          pip3 install -e .[test,geo,gpu,sglang] --no-deps && pip install transformers==4.52.3
+          pip3 install -e .[test,geo,gpu,sglang] --no-deps
+          pip install "transformers[hf_xet]==4.54.0"
       # Geo3k
       - name: Prepare GEO3K dataset
         run: |

diff --git a/.github/workflows/e2e_ppo_trainer_megatron_sglang.yml b/.github/workflows/e2e_ppo_trainer_megatron_sglang.yml
@@ -86,7 +86,7 @@ permissions:
   contents: read
 
 env:
-  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2"
+  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2"
   DYNAMIC_RUNNER_ENDPOINT: "https://sd10g3clalm04ug7alq90.apigateway-cn-beijing.volceapi.com/runner"
 
 jobs:

diff --git a/.github/workflows/e2e_ppo_trainer_megatron_vllm.yml b/.github/workflows/e2e_ppo_trainer_megatron_vllm.yml
@@ -85,7 +85,7 @@ permissions:
   contents: read
 
 env:
-  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2"
+  IMAGE: "verl-ci-cn-beijing.cr.volces.com/verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2"
   DYNAMIC_RUNNER_ENDPOINT: "https://sd10g3clalm04ug7alq90.apigateway-cn-beijing.volceapi.com/runner"
 
 jobs:
@@ -348,7 +348,6 @@ jobs:
       - name: Install the current repository
         run: |
           pip3 install --no-deps -e .[test]
-          pip3 install "transformers[hf_xet]<4.52.0"
       - name: Prepare Geo3k dataset
         run: |
           python3 examples/data_preprocess/geo3k.py

diff --git a/.github/workflows/e2e_spin.yml b/.github/workflows/e2e_spin.yml
@@ -68,7 +68,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/e2e_sppo.yml b/.github/workflows/e2e_sppo.yml
@@ -66,7 +66,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/gpu_unit_tests.yml b/.github/workflows/gpu_unit_tests.yml
@@ -80,7 +80,7 @@ jobs:
       NO_PROXY: "localhost,127.0.0.1"
       HF_HUB_ENABLE_HF_TRANSFER: 1
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/model.yml b/.github/workflows/model.yml
@@ -73,7 +73,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.4-sglang0.4.6.post5-vllm0.8.5-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
@@ -82,7 +82,7 @@ jobs:
       - name: Install the current repository and upgrade to latest transformers(4.54.0)/flash_attn, transformers 4.55.0 has strange behavior with model backward
         run: |
           pip3 install --no-deps -e .[test]
-          pip3 install --upgrade transformers==4.54.0
+          pip3 install --upgrade transformers
       - name: Running rmpad model tests on 8 L20 GPUs + flash_attn 2.5.8
         run: |
           pytest -s tests/models/test_transformer.py
@@ -95,6 +95,10 @@ jobs:
       - name: Running transformers ulysses tests on 8 L20 GPUs + latest transformers
         run: |
           torchrun --nproc_per_node=8 -m pytest tests/models/test_transformers_ulysses.py
+      - name: Running transformers ulysses tests on 8 L20 GPUs + transformers 4.54.1
+        run: |
+          pip3 install transformers==4.54.1
+          torchrun --nproc_per_node=8 -m pytest tests/models/test_transformers_ulysses.py
       - name: Running transformers ulysses tests on 8 L20 GPUs + transformers 4.53.2
         run: |
           pip3 install transformers==4.53.2
@@ -119,7 +123,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:base-verl0.5-cu126-cudnn9.8-torch2.7.0-fa2.7.4
+      image: verlai/verl:base-verl0.5-cu126-cudnn9.8-torch2.7.1-fa2.7.4
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/sgl.yml b/.github/workflows/sgl.yml
@@ -90,7 +90,7 @@ jobs:
       NCCL_SHM_DISABLE: "1"
       NCCL_P2P_DISABLE: "1"
     container:
-      image: verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/.github/workflows/vllm.yml b/.github/workflows/vllm.yml
@@ -84,7 +84,7 @@ jobs:
       HF_ENDPOINT: "https://hf-mirror.com"
       HF_HUB_ENABLE_HF_TRANSFER: "0" # This is more stable
     container:
-      image: verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2
+      image: verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2
       options: --gpus all --shm-size=10g
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2

diff --git a/docker/README.md b/docker/README.md
@@ -14,9 +14,7 @@ The first two types of images are hosted on dockerhub [verlai/verl](https://hub.
 
 ## Base Image
 
-The stable base image is ``verlai/verl:base-verl0.4-cu124-cudnn9.8-torch2.6-fa2.7.4``. The installed package versions can be found from tags, and the Dockerfile can be found in ``verl[version]-[packages]/Dockerfile.base``.
-
-The base images for preview are ``verlai/verl:base-verl0.5-cu126-cudnn9.8-torch2.7.1-fa2.8.0`` and ``verlai/verl:base-verl0.5-preview-cu128-cudnn9.8-torch2.7.1-fa2.8.0`` with different CUDA versions.
+The stable base image is ``verlai/verl:base-verl0.5-cu126-cudnn9.8-torch2.7.1-fa2.7.4`` with different CUDA versions.
 
 The update of base image is not frequent, and the app image can be built on top of it without reinstalling base packages.
 
@@ -25,8 +23,8 @@ The update of base image is not frequent, and the app image can be built on top
 From this version, we divide images built for vLLM and SGLang as the divergence of dependent packages like FlashInfer.
 There are 2 types of application images available:
 
-- **vLLM with FSDP and Megatron**: ``verlai/verl:app-verl0.5-vllm0.9.1-mcore0.12.2-te2.2``
-- **SGLang with FSDP and Megatron**: ``verlai/verl:app-verl0.5-sglang0.4.9.post6-mcore0.12.2-te2.2``
+- **vLLM with FSDP and Megatron**: ``verlai/verl:app-verl0.5-transformers4.55.4-vllm0.10.0-mcore0.13.0-te2.2``
+- **SGLang with FSDP and Megatron**: `verlai/verl:app-verl0.5-transformers4.55.4-sglang0.4.9.post6-mcore0.13.0-te2.2`
 
 Docker images with Megatron backends are runnable with large language model like ``Qwen/Qwen3-235B-A22B``, ``deepseek-ai/DeepSeek-V3-0324`` post-training. Refer to the :doc:`Large Language Model Post-Training documentation<../perf/dpsk>` for more details.
 
@@ -54,7 +52,7 @@ docker start verl
 docker exec -it verl bash
 ```
 
-2.	If you use the images provided, you only need to install verl itself without dependencies:
+2. If you use the images provided, you only need to install verl itself without dependencies:
 
 ```sh
 # install the nightly version (recommended)

diff --git a/docker/verl0.5-cu126-torch2.7-fa2.7.4/Dockerfile.app.sglang0.4.8.mcore0.12 b/docker/verl0.5-cu126-torch2.7-fa2.7.4/Dockerfile.app.sglang0.4.8.mcore0.12
diff --git a/docker/verl0.5-cu126-torch2.7-fa2.7.4/Dockerfile.app.sglang0.4.9.mcore0.12 b/docker/verl0.5-cu126-torch2.7-fa2.7.4/Dockerfile.app.sglang0.4.9.mcore0.12
diff --git a/...ockerfile.app.sglang0.4.9.post6.mcore0.12 → ...ockerfile.app.sglang0.4.9.post6.mcore0.13 b/...ockerfile.app.sglang0.4.9.post6.mcore0.12 → ...ockerfile.app.sglang0.4.9.post6.mcore0.13
@@ -17,7 +17,7 @@ RUN pip install --resume-retries 999 --no-cache-dir --no-build-isolation flashin
 RUN pip install --resume-retries 999  --no-cache-dir --no-build-isolation "sglang[all]==0.4.9.post6"
 
 # Fix packages
-RUN pip install --no-cache-dir "tensordict==0.6.2" "transformers[hf_xet]==4.54.0" accelerate datasets peft hf-transfer \
+RUN pip install --no-cache-dir "tensordict==0.6.2" "transformers[hf_xet]==4.55.4" accelerate datasets peft hf-transfer \
     "numpy<2.0.0" "pyarrow>=19.0.1" pandas \
     ray[default] codetiming hydra-core pylatexenc qwen-vl-utils wandb dill pybind11 liger-kernel mathruler blobfile xgrammar \
     pytest py-spy pyext pre-commit ruff
@@ -31,7 +31,7 @@ RUN pip install --resume-retries 999 --no-cache-dir nvidia-cudnn-cu12==9.8.0.87
 RUN export NVTE_FRAMEWORK=pytorch && pip3 install --resume-retries 999 --no-deps --no-cache-dir --no-build-isolation git+https://github.com/NVIDIA/TransformerEngine.git@v2.2.1
 
 # Install Megatron-LM
-RUN pip3 install --no-deps --no-cache-dir --no-build-isolation git+https://github.com/NVIDIA/Megatron-LM.git@core_v0.12.2
+RUN pip3 install --no-deps --no-cache-dir --no-build-isolation git+https://github.com/NVIDIA/Megatron-LM.git@core_v0.13.0
 
 # Install mbridge
 RUN pip3 install --no-cache-dir mbridge