Multi Pin Bumps across PT/AO/tune/ET

install/install_requirements.sh

@@ -60,10 +60,10 @@ echo "Using pip executable: $PIP_EXECUTABLE"
 # NOTE: If a newly-fetched version of the executorch repo changes the value of
 # PYTORCH_NIGHTLY_VERSION, you should re-run this script to install the necessary
 # package versions.
-PYTORCH_NIGHTLY_VERSION=dev20241002
+PYTORCH_NIGHTLY_VERSION=dev20241112
 
 # Nightly version for torchvision
-VISION_NIGHTLY_VERSION=dev20241002
+VISION_NIGHTLY_VERSION=dev20241112
 
 # Nightly version for torchtune
 TUNE_NIGHTLY_VERSION=dev20241010

torchchat/cli/builder.py

@@ -402,6 +402,7 @@ def _load_checkpoint(builder_args: BuilderArgs):
                     os.path.join(builder_args.checkpoint_dir, cp_name),
                     map_location=builder_args.device,
                     mmap=True,
+                    weights_only=False,
                 )
             )
         checkpoint = {}
@@ -706,4 +707,4 @@ def tokenizer_setting_to_name(tiktoken: bool, tokenizers: bool) -> str:
         return "TikToken"
     if tokenizers:
         return "Tokenizers"
-    return "SentencePiece"
+    return "SentencePiece"

torchchat/distributed/checkpoint.py

@@ -96,6 +96,7 @@ def _load_checkpoints_from_storage(
         checkpoint_path,
         map_location=builder_args.device,
         mmap=True,
+        weight_only=False,
     )
 
 

torchchat/export.py

@@ -122,7 +122,7 @@ def export_for_server(
     from executorch.exir.tracer import Value
 
     from torch._export import capture_pre_autograd_graph
-    from torch.export import export, ExportedProgram
+    from torch.export import export_for_training, ExportedProgram
 
     from torchchat.model import apply_rotary_emb, Attention
     from torchchat.utils.build_utils import get_precision
@@ -238,7 +238,7 @@ def _to_core_aten(
             raise ValueError(
                 f"Expected passed in model to be an instance of fx.GraphModule, got {type(model)}"
             )
-        core_aten_ep = export(model, example_inputs, dynamic_shapes=dynamic_shapes)
+        core_aten_ep = export_for_training(model, example_inputs, dynamic_shapes=dynamic_shapes)
         if verbose:
             logging.info(f"Core ATen graph:\n{core_aten_ep.graph}")
         return core_aten_ep