changes after review

Author: Abdennacer-Badaoui

.github/workflows/matrix.json

@@ -126,11 +126,10 @@
     "dockerfile": "Dockerfile-intel"
   },
   {
-    "name": "rocm-mi300-mi325",
+    "name": "rocm-gfx942-gfx950",
     "imageNamePrefix": "rocm-",
     "runOn": "always",
     "sccache": true,
-    "extraBuildArgs": "GPU_ARCHS=gfx942",
     "grpc": true,
     "dockerfile": "Dockerfile-amd"
   }

Dockerfile-amd

@@ -1,4 +1,3 @@
-# ── Rust builder ─────────────────────────────────────────────────────────────
 FROM lukemathwalker/cargo-chef:latest-rust-1.92-bookworm AS chef
 WORKDIR /usr/src
 
@@ -58,7 +57,6 @@ RUN --mount=type=secret,id=actions_results_url,env=ACTIONS_RESULTS_URL \
     --mount=type=secret,id=actions_runtime_token,env=ACTIONS_RUNTIME_TOKEN \
     cargo build --release --bin text-embeddings-router -F grpc -F python --no-default-features && sccache -s
 
-# ── ROCm Python runtime ───────────────────────────────────────────────────────
 FROM rocm/pytorch:rocm7.2.2_ubuntu22.04_py3.10_pytorch_release_2.10.0 AS rocm
 
 LABEL maintainer="Hugging Face"
@@ -87,7 +85,6 @@ COPY backends/python/server/requirements-amd.txt backends/python/server/requirem
 RUN cd backends/python/server && \
     make install
 
-# ── Final images ──────────────────────────────────────────────────────────────
 FROM rocm AS grpc
 
 COPY --from=grpc-builder /usr/src/target/release/text-embeddings-router /usr/local/bin/text-embeddings-router

docs/source/en/amd_gpu.md

@@ -31,7 +31,7 @@ model=BAAI/bge-base-en-v1.5
 volume=$PWD/data  # share a volume to avoid re-downloading weights
 
 docker run \
-  --device /dev/kfd --device /dev/dri \
+  --device /dev/kfd --device /dev/dri/renderD128 \
   --group-add video \
   --ipc=host \
   -p 8080:80 \
@@ -44,10 +44,13 @@ docker run \
 Then test it:
 
 ```shell
-curl http://localhost:8080/embed \
-    -X POST \
+curl http://localhost:8080/v1/embeddings \
     -H 'Content-Type: application/json' \
-    -d '{"inputs": "What is Deep Learning?"}'
+    -d '{
+      "input": "What is Deep Learning?",
+      "model": "text-embeddings-inference",
+      "encoding_format": "float"
+    }'
 ```
 
 ---
@@ -59,7 +62,7 @@ If you prefer to build from source, use AMD's official ROCm PyTorch image as the
 ## Step 1: Start the container
 
 ```shell
-docker run -it --device=/dev/kfd --device=/dev/dri \
+docker run -it --device=/dev/kfd --device=/dev/dri/renderD128 \
   --group-add video --shm-size 8g \
   -v $PWD:/workspace \
   rocm/pytorch:latest bash
@@ -132,10 +135,13 @@ model=BAAI/bge-base-en-v1.5
 Once the server is ready, you can test it with a simple embed request:
 
 ```shell
-curl http://localhost:8080/embed \
-    -X POST \
+curl http://localhost:8080/v1/embeddings \
     -H 'Content-Type: application/json' \
-    -d '{"inputs": "What is Deep Learning?"}'
+    -d '{
+      "input": "What is Deep Learning?",
+      "model": "text-embeddings-inference",
+      "encoding_format": "float"
+    }'
 ```
 
 ## Verifying GPU detection