feat(wandb): logging and configuration improvements (#10826)

* feat: ensure unique artifact id

* feat: allow manual init

* fix: simplify reinit logic

* fix: no dropped value + immediate commits

* fix: wandb use in sagemaker

* docs: improve documenation and formatting

* fix: typos

* docs: improve formatting

examples/README.md

@@ -240,34 +240,11 @@ Whenever you use `Trainer` or `TFTrainer` classes, your losses, evaluation metri
 
 Advanced configuration is possible by setting environment variables:
 
-<table>
-  <thead>
-    <tr>
-      <th style="text-align:left">Environment Variables</th>
-      <th style="text-align:left">Options</th>
-    </tr>
-  </thead>
-  <tbody>
-    <tr>
-      <td style="text-align:left">WANDB_LOG_MODEL</td>
-      <td style="text-align:left">Log the model as artifact at the end of training (<b>false</b> by default)</td>
-    </tr>
-    <tr>
-      <td style="text-align:left">WANDB_WATCH</td>
-      <td style="text-align:left">
-        <ul>
-          <li><b>gradients</b> (default): Log histograms of the gradients</li>
-          <li><b>all</b>: Log histograms of gradients and parameters</li>
-          <li><b>false</b>: No gradient or parameter logging</li>
-        </ul>
-      </td>
-    </tr>
-    <tr>
-      <td style="text-align:left">WANDB_PROJECT</td>
-      <td style="text-align:left">Organize runs by project</td>
-    </tr>
-  </tbody>
-</table>
+| Environment Variable | Value |
+|---|---|
+| WANDB_LOG_MODEL | Log the model as artifact (log the model as artifact at the end of training (`false` by default) |
+| WANDB_WATCH | one of `gradients` (default) to log histograms of gradients, `all` to log histograms of both gradients and parameters, or `false` for no histogram logging |
+| WANDB_PROJECT | Organize runs by project |
 
 Set run names with `run_name` argument present in scripts or as part of `TrainingArguments`.
 

src/transformers/integrations.py

@@ -19,7 +19,6 @@ import io
 import json
 import numbers
 import os
-import re
 import tempfile
 from copy import deepcopy
 from pathlib import Path
@@ -559,20 +558,12 @@ class WandbCallback(TrainerCallback):
         if has_wandb:
             import wandb
 
-            wandb.ensure_configured()
-            if wandb.api.api_key is None:
-                has_wandb = False
-                logger.warning(
-                    "W&B installed but not logged in. Run `wandb login` or set the WANDB_API_KEY env variable."
-                )
-                self._wandb = None
-            else:
-                self._wandb = wandb
+            self._wandb = wandb
         self._initialized = False
         # log outputs
         self._log_model = os.getenv("WANDB_LOG_MODEL", "FALSE").upper() in ENV_VARS_TRUE_VALUES.union({"TRUE"})
 
-    def setup(self, args, state, model, reinit, **kwargs):
+    def setup(self, args, state, model, **kwargs):
         """
         Setup the optional Weights & Biases (`wandb`) integration.
 
@@ -581,7 +572,8 @@ class WandbCallback(TrainerCallback):
 
         Environment:
             WANDB_LOG_MODEL (:obj:`bool`, `optional`, defaults to :obj:`False`):
-                Whether or not to log model as artifact at the end of training.
+                Whether or not to log model as artifact at the end of training. Use along with
+                `TrainingArguments.load_best_model_at_end` to upload best model.
             WANDB_WATCH (:obj:`str`, `optional` defaults to :obj:`"gradients"`):
                 Can be :obj:`"gradients"`, :obj:`"all"` or :obj:`"false"`. Set to :obj:`"false"` to disable gradient
                 logging or :obj:`"all"` to log gradients and parameters.
@@ -610,13 +602,19 @@ class WandbCallback(TrainerCallback):
             else:
                 run_name = args.run_name
 
-            self._wandb.init(
-                project=os.getenv("WANDB_PROJECT", "huggingface"),
-                config=combined_dict,
-                name=run_name,
-                reinit=reinit,
-                **init_args,
-            )
+            if self._wandb.run is None:
+                self._wandb.init(
+                    project=os.getenv("WANDB_PROJECT", "huggingface"),
+                    name=run_name,
+                    **init_args,
+                )
+            # add config parameters (run may have been created manually)
+            self._wandb.config.update(combined_dict, allow_val_change=True)
+
+            # define default x-axis (for latest wandb versions)
+            if getattr(self._wandb, "define_metric", None):
+                self._wandb.define_metric("train/global_step")
+                self._wandb.define_metric("*", step_metric="train/global_step", step_sync=True)
 
             # keep track of model topology and gradients, unsupported on TPU
             if not is_torch_tpu_available() and os.getenv("WANDB_WATCH") != "false":
@@ -628,23 +626,20 @@ class WandbCallback(TrainerCallback):
         if self._wandb is None:
             return
         hp_search = state.is_hyper_param_search
-        if not self._initialized or hp_search:
-            self.setup(args, state, model, reinit=hp_search, **kwargs)
+        if hp_search:
+            self._wandb.finish()
+        if not self._initialized:
+            self.setup(args, state, model, **kwargs)
 
     def on_train_end(self, args, state, control, model=None, tokenizer=None, **kwargs):
         if self._wandb is None:
             return
-        # commit last step
-        if state.is_world_process_zero:
-            self._wandb.log({})
         if self._log_model and self._initialized and state.is_world_process_zero:
             from .trainer import Trainer
 
             fake_trainer = Trainer(args=args, model=model, tokenizer=tokenizer)
             with tempfile.TemporaryDirectory() as temp_dir:
                 fake_trainer.save_model(temp_dir)
-                # use run name and ensure it's a valid Artifact name
-                artifact_name = re.sub(r"[^a-zA-Z0-9_\.\-]", "", self._wandb.run.name)
                 metadata = (
                     {
                         k: v
@@ -657,7 +652,7 @@ class WandbCallback(TrainerCallback):
                         "train/total_floss": state.total_flos,
                     }
                 )
-                artifact = self._wandb.Artifact(name=f"run-{artifact_name}", type="model", metadata=metadata)
+                artifact = self._wandb.Artifact(name=f"model-{self._wandb.run.id}", type="model", metadata=metadata)
                 for f in Path(temp_dir).glob("*"):
                     if f.is_file():
                         with artifact.new_file(f.name, mode="wb") as fa:
@@ -668,10 +663,10 @@ class WandbCallback(TrainerCallback):
         if self._wandb is None:
             return
         if not self._initialized:
-            self.setup(args, state, model, reinit=False)
+            self.setup(args, state, model)
         if state.is_world_process_zero:
             logs = rewrite_logs(logs)
-            self._wandb.log(logs, step=state.global_step)
+            self._wandb.log({**logs, "train/global_step": state.global_step})
 
 
 class CometCallback(TrainerCallback):