Add outputs

llmfoundry/callbacks/eval_output_logging_callback.py

@@ -47,22 +47,25 @@ def eval_batch_end(self, state: State, logger: Logger) -> None:
                                           state.metric_outputs,
                                       )
 
-        # If batch mode is not generate, outputs will be logits
+
         if state.batch.get('mode') == 'generate':
             # Outputs are already detokenized
             logging_dict['outputs'] = state.outputs
+        elif isinstance(state.outputs, torch.Tensor):
+            # If batch mode is not generate, outputs will be logits
+            logging_dict['outputs'] = state.outputs.argmax(dim=-1)
 
         input_ids = state.batch['input_ids']
         logged_input = []
         assert state.dataloader is not None
+        dataset = state.dataloader.dataset  # pyright: ignore[reportGeneralTypeIssues]
+        tokenizer = dataset.tokenizer  # pyright: ignore[reportGeneralTypeIssues]
+        pad_token_id = getattr(dataset, 'pad_tok_id', dataset.tokenizer.pad_token_id)
 
         # Depad and decode input_ids
         for input_list in input_ids.tolist():
-            dataset = state.dataloader.dataset  # pyright: ignore[reportGeneralTypeIssues]
-            depadded_input = [
-                tok for tok in input_list if tok != dataset.tokenizer.pad_token_id
-            ]
-            logged_input.append(dataset.tokenizer.decode(depadded_input))
+            depadded_input = [tok for tok in input_list if tok != pad_token_id]
+            logged_input.append(tokenizer.decode(depadded_input))
         logging_dict['input'] = logged_input
 
         # Log token indices if toggled
@@ -85,18 +88,12 @@ def eval_batch_end(self, state: State, logger: Logger) -> None:
         for key, value in logging_dict.items():
             # All types in list are the same
             if isinstance(value[0], torch.Tensor):
-                logging_dict[key] = [
-                    state.dataloader.dataset. # pyright: ignore[reportGeneralTypeIssues]
-                    tokenizer.decode(  # pyright: ignore[reportGeneralTypeIssues]
-                        t,
-                    ) for t in value
-                ]
+                logging_dict[key] = [tokenizer.decode(t) for t in value]
             elif isinstance(value[0], list):
                 if isinstance(value[0][0], torch.Tensor):
-                    tokenizer = state.dataloader.dataset.tokenizer  # pyright: ignore[reportGeneralTypeIssues]
-                    logging_dict[key] = [[
-                        tokenizer.decode(choice) for choice in t
-                    ] for t in value]
+                    logging_dict[key] = [
+                        [tokenizer.decode(choice) for choice in t] for t in value
+                    ]
 
         # Convert logging_dict from kv pairs of column name and column values to a list of rows
         # Example: