inc only working with quantized models

huggingface · Jul 1, 2024 · b040d78 · b040d78
1 parent dd02f26
commit b040d78
Show file tree

Hide file tree

Showing 4 changed files with 6 additions and 1 deletion.
diff --git a/tests/configs/_inc_quant_.yaml b/tests/configs/_inc_quant_.yaml
@@ -0,0 +1,3 @@
+backend:
+  ptq_quantization: true
+  calibration: true
diff --git a/tests/configs/_text_decoders_.yaml b/tests/configs/_text_decoders_.yaml
@@ -3,4 +3,4 @@ hydra:
   sweeper:
     params:
       backend.task: text-generation
-      backend.model: hf-internal-testing/tiny-random-gpt2,IlyasMoutawwakil/tiny-random-llama
+      backend.model: hf-internal-testing/tiny-random-GPT2LMHeadModel,hf-internal-testing/tiny-random-LlamaForCausalLM
diff --git a/...ence_neural_compressor_text_decoders.yaml → ...l_compressor_inc_quant_text_decoders.yaml b/...ence_neural_compressor_text_decoders.yaml → ...l_compressor_inc_quant_text_decoders.yaml
@@ -3,6 +3,7 @@ defaults:
   - _base_ # inherits from base config
   - _cpu_ # inherits from cpu config
   - _inference_ # inherits from inference config
+  - _inc_quant_ # inherits from incremental quantization config
   - _text_decoders_ # inherits from text decoders config
   - _no_weights_ # inherits from no weights config
   - _self_ # hydra 1.1 compatibility

diff --git a/...ence_neural_compressor_text_encoders.yaml → ...l_compressor_inc_quant_text_encoders.yaml b/...ence_neural_compressor_text_encoders.yaml → ...l_compressor_inc_quant_text_encoders.yaml
@@ -3,6 +3,7 @@ defaults:
   - _base_ # inherits from base config
   - _cpu_ # inherits from cpu config
   - _inference_ # inherits from inference config
+  - _inc_quant_ # inherits from incremental quantization config
   - _text_encoders_ # inherits from text encoders config
   - _no_weights_ # inherits from no weights config
   - _self_ # hydra 1.1 compatibility