0.0.17

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.so filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -1 +1,2 @@
 apps.json
+kaggle.json

Dockerfile

@@ -1,13 +1,13 @@
-# build with: docker build . --tag sctg/roco-idefics3:0.0.15 --tag sctg/roco-idefics3:latest --push
+# build with: docker build . --tag sctg/roco-idefics3:0.0.17 --tag sctg/roco-idefics3:latest --push
 # run with 
-# docker run --gpus all --user=42420:42420 -p 7000-8000:7000-8000 -p 8080:8080 -e HF_TOKEN=hf_TOKEN -it sctg/roco-idefics3:0.0.15 bash -i /start.sh sleep infinity
-# docker run --gpus all --user=42420:42420 -p 7000-8000:7000-8000 -p 8080:8080 -it sctg/roco-idefics3:0.0.15 bash -i /start.sh python /learn.py hf_...
+# docker run --gpus all --user=42420:42420 -p 7000-8000:7000-8000 -p 8080:8080 -e WANDB_API_KEY=wdkfjzfjz -e HF_TOKEN=hf_TOKEN -it sctg/roco-idefics3:0.0.17 bash -i /start.sh sleep infinity
+# docker run --gpus all --user=42420:42420 -p 7000-8000:7000-8000 -p 8080:8080 -it sctg/roco-idefics3:0.0.17 bash -i /start.sh python /learn.py hf_...
 FROM cloudflare/cloudflared:latest as cloudflared
-FROM nvidia/cuda:12.6.2-devel-ubuntu22.04
+FROM nvidia/cuda:12.6.2-cudnn-devel-ubuntu22.04
 ARG NODE_MAJOR="20"
 RUN /usr/sbin/addgroup --gid 42420 ovh
 RUN /usr/sbin/useradd -u 42420 --gid 42420 -m -d /workspace -s /bin/bash ovh
-RUN apt update -y && apt-get install -y curl git git-lfs screen sudo \
+RUN apt update -y && apt-get install -y cmake curl git git-lfs screen sudo \
       && mkdir -p /etc/apt/keyrings \
       && curl -fsSL https://deb.nodesource.com/gpgkey/nodesource-repo.gpg.key | gpg --dearmor -o /etc/apt/keyrings/nodesource.gpg \
       && echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_$NODE_MAJOR.x nodistro main" | tee /etc/apt/sources.list.d/nodesource.list \
@@ -27,6 +27,7 @@ RUN . /workspace/.miniconda3/bin/activate && conda init --all
 RUN . /workspace/.miniconda3/bin/activate \
     && pip install -U "safetensors>=0.4.5" \
     && pip install -U tensorflow \
+    && pip install -U tf-keras \
     && pip install -U 'https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_multi-backend-refactor/bitsandbytes-0.44.1.dev0-py3-none-manylinux_2_24_x86_64.whl' \
     && pip install -U git+https://github.com/huggingface/transformers.git\
     && pip install huggingface_hub[cli] accelerate datasets peft\
@@ -39,12 +40,16 @@ RUN . /workspace/.miniconda3/bin/activate \
     && pip install unsloth\
     && pip install gradio \
     && pip uninstall unsloth -y \
-    && pip install --upgrade --no-cache-dir --no-deps git+https://github.com/unslothai/unsloth.git
+    && pip install --upgrade --no-cache-dir --no-deps git+https://github.com/unslothai/unsloth.git \
+    && pip install wandb
+    
+COPY --chmod=777 libbitsandbytes_cuda124.so /workspace/.miniconda3/lib/python3.12/site-packages/bitsandbytes/libbitsandbytes_cuda124.so
 
 RUN . /workspace/.miniconda3/bin/activate && conda install -y jupyter
 RUN rm -f /workspace/miniconda.sh
 RUN mkdir -p /workspace/.config/github-copilot/
 COPY apps.json.enc /workspace/.config/github-copilot/apps.json.enc
+COPY kaggle.json.enc /workspace/.config/kaggle.json.enc
 COPY CFTOKEN.enc /workspace/.config/CFTOKEN.enc
 # Mandatory to run the jobs in rootless mode
 # USER root

kaggle.json.enc

@@ -0,0 +1,2 @@
+U2FsdGVkX1+XKfjZhSpObwI5HP7f0R/U/8Paum3H41R7B3HImmd1rUJNeG4x1zFJ
+HqBGdGRcJiWwKkMVOCQyFQrNJTKKOFt/wDk1FWqEoQ8=

libbitsandbytes_cuda124.so

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41c9698af520efd7212d164e394aadb42d45b5fdee0b36ed31ae25422ac29c30
+size 34242400

start.sh

@@ -37,6 +37,14 @@ if [ -f /workspace/.config/CFTOKEN.enc ]; then
   /usr/local/bin/cloudflared --pidfile /tmp/cf.pid  --autoupdate-freq 24h0m0s tunnel run --token $CFTOKEN &
 fi
 
+# Decode kaggle.json.enc file
+# encoded with: openssl aes-256-cbc -base64 -md sha256 -pass pass:"$HF_TOKEN" -in kaggle.json -out kaggle.json.enc
+# decode with: openssl aes-256-cbc -a -d -md sha256 -pass pass:"$HF_TOKEN" -in kaggle.json.enc -out kaggle.json
+if [ -f /workspace/.config/kaggle.json.enc ]; then
+  mkdir -p $HOME/.kaggle
+  openssl aes-256-cbc -a -d -md sha256 -pass pass:"$HF_TOKEN" -in /workspace/.config/kaggle.json.enc -out /workspace/.kaggle/kaggle.json
+fi
+
 git clone https://huggingface.co/eltorio/IDEFICS3_ROCOv2
 git config --global user.email "me@hg.co"
 git config --global user.name "me@hg.co"