diff --git a/models/turbine_models/custom_models/llm_cmd_opts.py b/models/turbine_models/custom_models/llm_cmd_opts.py index 3415acac..b0396175 100644 --- a/models/turbine_models/custom_models/llm_cmd_opts.py +++ b/models/turbine_models/custom_models/llm_cmd_opts.py @@ -44,8 +44,8 @@ def is_valid_file(arg): ) ############################################################################## -# SDXL Inference Options -# These options are used to control runtime parameters for SDXL inference. +# Runtime Options +# These options are used to control runtime parameters. ############################################################################## p.add_argument( diff --git a/models/turbine_models/custom_models/stateless_llama.py b/models/turbine_models/custom_models/stateless_llama.py index 5ed4e0e2..30985f4f 100644 --- a/models/turbine_models/custom_models/stateless_llama.py +++ b/models/turbine_models/custom_models/stateless_llama.py @@ -758,6 +758,9 @@ def format_out(results): llama = StatelessLlama( args.hf_model_name, args.precision, + args.max_length, + args.batch_size, + args.num_inference_steps, args.device, args.iree_target_triple, flags,