use older tgi version

huggingface · Dec 16, 2024 · 3b0138b · 3b0138b
1 parent bd54289
commit 3b0138b
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 1 deletion.
diff --git a/examples/cuda_tgi_llama.yaml b/examples/cuda_tgi_llama.yaml
@@ -18,6 +18,7 @@ backend:
   cuda_graphs: 0 # remove for better perf but bigger memory footprint
   no_weights: true
   model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
+  image: ghcr.io/huggingface/text-generation-inference:2.4.1
 
 scenario:
   input_shapes:

diff --git a/tests/configs/_no_weights_.yaml b/tests/configs/_no_weights_.yaml
@@ -2,4 +2,4 @@ hydra:
   mode: MULTIRUN
   sweeper:
     params:
-      backend.no_weights: true
+      backend.no_weights: true,false
diff --git a/tests/configs/cpu_inference_py_txi_gpt2.yaml b/tests/configs/cpu_inference_py_txi_gpt2.yaml
@@ -12,3 +12,4 @@ name: cpu_inference_py_txi_gpt2
 
 backend:
   cuda_graphs: 0
+  image: ghcr.io/huggingface/text-generation-inference:2.4.1
diff --git a/tests/configs/cuda_inference_py_txi_gpt2.yaml b/tests/configs/cuda_inference_py_txi_gpt2.yaml
@@ -12,3 +12,4 @@ name: cuda_inference_py_txi_gpt2
 
 backend:
   cuda_graphs: 0
+  image: ghcr.io/huggingface/text-generation-inference:2.4.1
Original file line number	Diff line number	Diff line change
Expand Up		@@ -12,3 +12,4 @@ name: cpu_inference_py_txi_gpt2

		backend:
		cuda_graphs: 0
		image: ghcr.io/huggingface/text-generation-inference:2.4.1
Original file line number	Diff line number	Diff line change
Expand Up		@@ -12,3 +12,4 @@ name: cuda_inference_py_txi_gpt2

		backend:
		cuda_graphs: 0
		image: ghcr.io/huggingface/text-generation-inference:2.4.1