Merge branch 'huggingface:main' into minimal-structure

src/peft/peft_model.py

@@ -155,7 +155,8 @@ def from_pretrained(cls, model, model_id, **kwargs):
                     f"Please check that the file {WEIGHTS_NAME} is present at {model_id}."
                 )
 
-        adapters_weights = torch.load(filename)
+        adapters_weights = torch.load(
+            filename, map_location=torch.device("cuda" if torch.cuda.is_available() else "cpu"))
         # load the weights into the model
         model = set_peft_model_state_dict(model, adapters_weights)
         if getattr(model, "hf_device_map", None) is not None:
@@ -266,7 +267,12 @@ def print_trainable_parameters(self):
         trainable_params = 0
         all_param = 0
         for _, param in self.named_parameters():
-            all_param += param.numel()
+            num_params = param.numel()
+            # if using DS Zero 3 and the weights are initialized empty
+            if num_params == 0 and hasattr(param, "ds_numel"):
+                num_params = param.ds_numel
+
+            all_param += num_params
             if param.requires_grad:
                 trainable_params += param.numel()
         print(

src/peft/utils/other.py

@@ -30,7 +30,9 @@ def bloom_model_postprocess_past_key_value(past_key_values):
     return tuple(zip(keys, values))
 
 
-def prepare_model_for_int8_training(model, output_embedding_layer_name="lm_head"):
+def prepare_model_for_int8_training(
+    model, output_embedding_layer_name="lm_head", use_gradient_checkpointing=True, layer_norm_names=["layer_norm"]
+):
     r"""
     This method wrapps the entire protocol for preparing a model before running a training. This includes:
         1- Cast the layernorm in fp32 2- making output embedding layer require grads 3- Add the upcasting of the lm
@@ -48,20 +50,20 @@ def prepare_model_for_int8_training(model, output_embedding_layer_name="lm_head"
 
         if loaded_in_8bit:
             # cast layer norm in fp32 for stability for 8bit models
-            if param.ndim == 1 and "layer_norm" in name:
+            if param.ndim == 1 and any(layer_norm_name in name for layer_norm_name in layer_norm_names):
                 param.data = param.data.to(torch.float32)
 
-    # For backward compatibility
-    if hasattr(model, "enable_input_require_grads"):
-        model.enable_input_require_grads()
-    else:
+    if loaded_in_8bit and use_gradient_checkpointing:
+        # For backward compatibility
+        if hasattr(model, "enable_input_require_grads"):
+            model.enable_input_require_grads()
+        else:
 
-        def make_inputs_require_grad(module, input, output):
-            output.requires_grad_(True)
+            def make_inputs_require_grad(module, input, output):
+                output.requires_grad_(True)
 
-        model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
+            model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)
 
-    if loaded_in_8bit:
         # enable gradient checkpointing for memory efficiency
         model.gradient_checkpointing_enable()