add new kernels to engine

Author: Qing Lan

engines/python/setup/djl_python/properties_manager/hf_properties.py

@@ -24,8 +24,10 @@ class HFQuantizeMethods(str, Enum):
     awq = 'awq'
     deepspeedfp = 'deepspeedfp'
     fp8 = 'fp8'
+    fbgemm_fp8 = 'fbgemm_fp8'
     gptq_marlin = 'gptq_marlin'
     gptq_marlin_24 = 'gptq_marlin_24'
+    awq_marlin = 'awq_marlin'
     marlin = 'marlin'
     squeezellm = 'squeezellm'
 

engines/python/setup/djl_python/properties_manager/lmi_dist_rb_properties.py

@@ -22,9 +22,11 @@ class LmiDistQuantizeMethods(str, Enum):
     awq = 'awq'
     deepspeedfp = 'deepspeedfp'
     fp8 = 'fp8'
+    fbgemm_fp8 = 'fbgemm_fp8'
     gptq = 'gptq'
     gptq_marlin = 'gptq_marlin'
     gptq_marlin_24 = 'gptq_marlin_24'
+    awq_marlin = 'awq_marlin'
     marlin = 'marlin'
     squeezellm = 'squeezellm'
 

engines/python/setup/djl_python/properties_manager/vllm_rb_properties.py

@@ -22,9 +22,11 @@ class VllmQuantizeMethods(str, Enum):
     awq = 'awq'
     deepspeedfp = 'deepspeedfp'
     fp8 = 'fp8'
+    fbgemm_fp8 = 'fbgemm_fp8'
     gptq = 'gptq'
     gptq_marlin = 'gptq_marlin'
     gptq_marlin_24 = 'gptq_marlin_24'
+    awq_marlin = 'awq_marlin'
     marlin = 'marlin'
     squeezellm = 'squeezellm'