multigpu longer timeout

cicd/multigpu.py

@@ -68,7 +68,7 @@ def run_cmd(cmd: str, run_folder: str):
 @app.function(
     image=cicd_image,
     gpu=GPU_CONFIG,
-    timeout=60 * 60,
+    timeout=90 * 60,
     cpu=8.0,
     memory=131072 * N_GPUS,
     volumes=VOLUME_CONFIG,

docs/multimodal.qmd

@@ -164,7 +164,7 @@ Here is an example of a multi-modal dataset:
         {
             "role": "user",
             "content": [
-                {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
+                {"type": "image", "image": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
                 {"type": "text", "text": "Describe this image in detail."}
             ]
         },

requirements.txt

@@ -12,7 +12,7 @@ liger-kernel==0.5.6
 packaging==23.2
 
 peft==0.15.1
-transformers==4.51.0
+transformers==4.51.1
 tokenizers>=0.21.1
 accelerate==1.6.0
 datasets==3.5.0

src/axolotl/integrations/liger/README.md

@@ -29,7 +29,6 @@ liger_fused_linear_cross_entropy: true
 - granite
 - jamba
 - llama
-- llama4 (partial support, no support for FLCE yet)
 - mistral
 - mixtral
 - mllama