Add deepseek-r1 examples

examples/.dstack-task.yml

@@ -0,0 +1,15 @@
+type: dev-environment
+# The name is optional, if not specified, generated randomly
+name: vscode
+
+python: "3.11"
+# Uncomment to use a custom Docker image
+#image: dstackai/base:py3.13-0.6-cuda-12.1
+ide: vscode
+
+
+# Uncomment to leverage spot instances
+#spot_policy: auto
+
+resources:
+  gpu: 24GB

examples/.dstack.yml

@@ -1,15 +1,35 @@
-type: dev-environment
-# The name is optional, if not specified, generated randomly
-name: vscode
+type: service
+name: llama31
 
 python: "3.11"
-# Uncomment to use a custom Docker image
-#image: dstackai/base:py3.13-0.6-cuda-12.1
+env:
+  - HF_TOKEN
+  - MODEL_ID=meta-llama/Llama-3.2-1B
+  - MAX_MODEL_LEN=4096
+commands:
+  - pip install vllm
+  - curl -o simple_chat_template.jinja https://github.com/Bihan/vllm/blob/main/examples/simple_chat_template.jinja
+  - vllm serve $MODEL_ID
+    --max-model-len $MAX_MODEL_LEN
+    --chat-template simple_chat_template.jinja
 
-ide: vscode
+auth: false
 
-# Use either spot or on-demand instances
-spot_policy: auto
+port: 8000
+# Register the model
+model:
+  name: meta-llama/Llama-3.2-1B
+  type: chat
+  format: openai
+
+# Uncomment to leverage spot instances
+#spot_policy: auto
+
+# Uncomment to cache downloaded models
+#volumes:
+#  - /root/.cache/huggingface/hub:/root/.cache/huggingface/hub
 
 resources:
   gpu: 24GB
+  # Uncomment if using multiple GPUs
+  #shm_size: 24GB

examples/llms/deepseek/sglang/amd/.dstack.yml

@@ -0,0 +1,21 @@
+type: service
+name: deepseek-r1-amd
+
+image: lmsysorg/sglang:v0.4.1.post4-rocm620
+env:
+  - MODEL_ID=deepseek-ai/DeepSeek-R1-Distill-Llama-70B
+commands:
+  - python3 -m sglang.launch_server
+    --model-path $MODEL_ID
+    --port 8000
+    --trust-remote-code
+
+port: 8000
+model:
+  name: deepseek-ai/DeepSeek-R1-Distill-Llama-70B
+  type: chat
+  format: openai
+
+resources:
+    gpu: mi300x
+    disk: 300Gb

examples/llms/deepseek/sglang/nvidia/.dstack.yml

@@ -0,0 +1,22 @@
+type: service
+name: deepseek-r1-nvidia
+
+image: lmsysorg/sglang:latest
+env:
+  - MODEL_ID=deepseek-ai/DeepSeek-R1-Distill-Llama-8B
+commands:
+  - python3 -m sglang.launch_server
+    --model-path $MODEL_ID
+    --port 8000
+    --trust-remote-code
+
+port: 8000
+
+model:
+  name: deepseek-ai/DeepSeek-R1-Distill-Llama-8B
+  type: chat
+  format: openai
+
+
+resources:
+    gpu: 24GB

examples/llms/deepseek/vllm/amd/.dstack.yml

@@ -0,0 +1,24 @@
+type: service
+name: deepseek-r1-amd
+
+image: rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
+env:
+  - MODEL_ID=deepseek-ai/DeepSeek-R1-Distill-Llama-70B
+  - MAX_MODEL_LEN=4096
+
+commands:
+  - pip install vllm
+  - vllm serve $MODEL_ID
+    --max-model-len $MAX_MODEL_LEN
+
+port: 8000
+
+model:
+  name: deepseek-ai/DeepSeek-R1-Distill-Llama-70B
+  type: chat
+  format: openai
+
+
+resources:
+    gpu: mi300x
+    disk: 300Gb

examples/llms/deepseek/vllm/nvidia/.dstack.yml

@@ -0,0 +1,21 @@
+type: service
+name: deepseek-r1-nvidia
+
+image: vllm/vllm-openai:latest
+env:
+  - MODEL_ID=deepseek-ai/DeepSeek-R1-Distill-Llama-8B
+  - MAX_MODEL_LEN=4096
+commands:
+  - pip install vllm
+  - vllm serve $MODEL_ID
+    --max-model-len $MAX_MODEL_LEN
+
+port: 8000
+
+model:
+  name: deepseek-ai/DeepSeek-R1-Distill-Llama-8B
+  type: chat
+  format: openai
+
+resources:
+    gpu: 24GB