Fix infernece & typo (#602)

ppdiffusers/ppdiffusers/models/lora.py

@@ -304,7 +304,7 @@ def __init__(self, *args, lora_layer: Optional[LoRAConv2dLayer] = None, **kwargs
         self.out_channels = self._out_channels
         self.kernel_size = self._kernel_size
         self.lora_layer = lora_layer
-        self.data_format = kwargs["data_format"]
+        self.data_format = kwargs.get("data_format", "NCHW")
 
     def set_lora_layer(self, lora_layer: Optional[LoRAConv2dLayer]):
         self.lora_layer = lora_layer
@@ -366,11 +366,25 @@ def forward(self, hidden_states: paddle.Tensor, scale: float = 1.0) -> paddle.Te
             # make sure to the functional Conv2D function as otherwise torch.compile's graph will break
             # see: https://github.com/huggingface/diffusers/pull/4315
             return nn.functional.conv2d(
-                hidden_states, self.weight, self.bias, self._stride, self._padding, self._dilation, self._groups, data_format=self.data_format,
+                hidden_states,
+                self.weight,
+                self.bias,
+                self._stride,
+                self._padding,
+                self._dilation,
+                self._groups,
+                data_format=self.data_format,
             )
         else:
             original_outputs = nn.functional.conv2d(
-                hidden_states, self.weight, self.bias, self._stride, self._padding, self._dilation, self._groups, data_format=self.data_format,
+                hidden_states,
+                self.weight,
+                self.bias,
+                self._stride,
+                self._padding,
+                self._dilation,
+                self._groups,
+                data_format=self.data_format,
             )
             return original_outputs + (scale * self.lora_layer(hidden_states))
 

ppdiffusers/ppdiffusers/pipelines/kandinsky/text_encoder.py

@@ -94,5 +94,6 @@ def __init__(self, config, *args, **kwargs):
 
     def forward(self, input_ids, attention_mask):
         embs = self.transformer(input_ids=input_ids, attention_mask=attention_mask)[0]
+        attention_mask = attention_mask.cast(embs.dtype)
         embs2 = (embs * attention_mask.unsqueeze(2)).sum(axis=1) / attention_mask.sum(axis=1)[:, None]
         return self.LinearTransformation(embs2), embs

ppdiffusers/ppdiffusers/pipelines/stable_diffusion_xl/pipeline_paddleinfer_stable_diffusion_xl.py

@@ -12,6 +12,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+import os
 from typing import Any, Callable, Dict, List, Optional, Tuple, Union
 
 import paddle
@@ -461,8 +462,11 @@ def __call__(
 
                 # predict the noise residual
                 noise_pred = self.unet(**unet_inputs)[0]
-                if str(os.environ.get("FLAGS_model_return_data")).lower() in ("true", "1")::
-                    print(f"StableDiffusion infer: step {i+1} , origin output {noise_pred.abs().numpy().mean()} ", flush=True)
+                if str(os.environ.get("FLAGS_model_return_data")).lower() in ("true", "1"):
+                    print(
+                        f"StableDiffusion infer: step {i+1} , origin output {noise_pred.abs().numpy().mean()} ",
+                        flush=True,
+                    )
 
                 # perform guidance
                 if do_classifier_free_guidance:

ppdiffusers/ppdiffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero.py

@@ -237,7 +237,9 @@ def warp_single_latent(latent, reference_flow):
     coords_t0 = F.interpolate(coords_t0, size=(h, w), mode="bilinear")
     coords_t0 = paddle.transpose(coords_t0, (0, 2, 3, 1))
 
-    warped = grid_sample(latent, coords_t0, mode="nearest", padding_mode="reflection")
+    warped = grid_sample(
+        latent.cast("float32").cpu(), coords_t0.cast("float32").cpu(), mode="nearest", padding_mode="reflection"
+    )
     return warped
 
 

ppdiffusers/ppdiffusers/pipelines/unidiffuser/modeling_text_decoder.py

@@ -335,7 +335,7 @@ def generate_beam(
                 logits[is_stopped, 0] = 0
                 scores_sum = scores[:, None] + logits
                 seq_lengths[~is_stopped] += 1
-                scores_sum_average = scores_sum / seq_lengths[:, None]
+                scores_sum_average = scores_sum / seq_lengths[:, None].cast(scores_sum.dtype)
                 scores_sum_average, next_tokens = scores_sum_average.reshape([-1]).topk(beam_size, -1)
                 next_tokens_source = next_tokens // scores_sum.shape[1]
                 seq_lengths = seq_lengths[next_tokens_source]
@@ -344,7 +344,7 @@ def generate_beam(
                 tokens = tokens[next_tokens_source]
                 tokens = paddle.concat(x=(tokens, next_tokens), axis=1)
                 generated = generated[next_tokens_source]
-                scores = scores_sum_average * seq_lengths
+                scores = scores_sum_average * seq_lengths.cast(scores_sum_average.dtype)
                 is_stopped = paddle.cast(is_stopped, "int32")  # TODO: nf
                 is_stopped = is_stopped[next_tokens_source]
                 is_stopped = paddle.cast(is_stopped, "bool")
@@ -357,7 +357,7 @@ def generate_beam(
             if is_stopped.astype("bool").all():
                 break
 
-        scores = scores / seq_lengths
+        scores = scores / seq_lengths.cast(scores.dtype)
         order = scores.argsort(descending=True)
         # tokens tensors are already padded to max_seq_length
         output_texts = [tokens[i] for i in order]

ppdiffusers/ppdiffusers/utils/download_utils.py

@@ -244,7 +244,7 @@ def _resumable_file_manager():
             )
 
         logger.info("storing %s in cache at %s", url_to_download, file_path)
-        _chmod_and_move(temp_file.name, file_path)
+        _chmod_and_move(temp_file.name, Path(file_path))
     try:
         os.remove(lock_path)
     except OSError:
@@ -380,7 +380,7 @@ def _resumable_file_manager():
             )
 
         logger.info("storing %s in cache at %s", url_to_download, blob_path)
-        _chmod_and_move(temp_file.name, blob_path)
+        _chmod_and_move(temp_file.name, Path(blob_path))
     try:
         os.remove(lock_path)
     except OSError: