Add test documentation for AMD CPU Docker image

Author: randomm

TEST_AMD_IMAGE.md

@@ -0,0 +1,63 @@
+# Testing AMD CPU Compatible Docker Image
+
+This document explains how to test the AMD CPU compatible Docker image built by GitHub Actions.
+
+## Image Location
+
+The image will be available at:
+```
+ghcr.io/randomm/text-embeddings-inference/text-embeddings-inference-cpu-amd:fix-dockerfile-issue-7e9217e
+```
+
+## Quick Test on AMD Server
+
+1. **Pull the image:**
+   ```bash
+   docker pull ghcr.io/randomm/text-embeddings-inference/text-embeddings-inference-cpu-amd:fix-dockerfile-issue-7e9217e
+   ```
+
+2. **Run with Qwen3 model (the one that was failing):**
+   ```bash
+   model=Qwen/Qwen3-Embedding-0.6B
+   volume=$PWD/data
+   
+   docker run --rm -p 8080:80 -v $volume:/data \
+     ghcr.io/randomm/text-embeddings-inference/text-embeddings-inference-cpu-amd:fix-dockerfile-issue-7e9217e \
+     --model-id $model
+   ```
+
+3. **Test the embedding endpoint:**
+   ```bash
+   curl 127.0.0.1:8080/embed \
+     -X POST \
+     -d '{"inputs":"What is Deep Learning?"}' \
+     -H 'Content-Type: application/json'
+   ```
+
+## Expected Results
+
+✅ **Success indicators:**
+- No Intel MKL SGEMM errors in logs
+- Model loads successfully
+- Embedding requests return valid responses
+- No "Parameter 8/13 was incorrect" errors
+
+❌ **Previous error (should be gone):**
+```
+Intel MKL ERROR: Parameter 8 was incorrect on entry to SGEMM
+Intel MKL ERROR: Parameter 13 was incorrect on entry to SGEMM
+```
+
+## Key Differences from Original Image
+
+- **No Intel MKL**: Uses generic BLAS libraries
+- **AMD Compatible**: Works on AMD processors without CPU vendor detection issues
+- **Same Performance**: Should maintain similar performance characteristics
+- **Broader Compatibility**: Works on both Intel and AMD CPUs
+
+## Build Status
+
+Check the GitHub Actions build status at:
+https://github.com/randomm/text-embeddings-inference/actions
+
+The workflow "Build CPU AMD Docker Image" should show a successful build.