Sync eval changes in OLMo/ladder-1xC to here

CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 
 - Added support for tensor parallelism. See the `TransformerConfig` class for usage.
+- Added more downstream tasks from the model ladder.
 - Added `io.copy_dir()` function.
 - Added new LR schedulers: `LinearWithWarmup`, `InvSqrtWithWarmup`, `ConstantWithWarmup`, `SequentialScheduler`.
 

pyproject.toml

@@ -22,7 +22,7 @@ dependencies = [
     "omegaconf",
     "safetensors",
     "importlib_resources",
-    "ai2-olmo-eval==0.2.0",
+    "ai2-olmo-eval==0.5.0",
 ]
 
 [project.urls]

src/olmo_core/train/callbacks/evaluator_callback.py

@@ -99,7 +99,6 @@ def post_step(self):
             metrics = []
             with cuda_sync_debug_mode(0):
                 for name, value in evaluator.compute_metrics().items():
-                    value = value.item()
                     metrics.append(f"    {name}={format_float(value)}")
                     self.trainer.record_metric(f"eval/{evaluator.name}/{name}", value)
             log.info("Eval metrics:\n" + "\n".join(metrics))
@@ -161,6 +160,8 @@ class DownstreamEvaluator(Evaluator):
         "pmi_dc": "PMI-DC accuracy",
         "ce_loss": "CE loss",
         "bpb": "BPB",
+        "soft": "soft loss",
+        "soft_log": "log soft loss",
     }
 
     def __init__(
@@ -184,13 +185,14 @@ def __init__(
         if is_distributed():
             sampler = DistributedSampler(
                 self.task,  # type: ignore
-                drop_last=True,
+                drop_last=False,
                 shuffle=False,
                 num_replicas=get_world_size(dp_process_group),
                 rank=get_rank(dp_process_group),
             )
 
         rank_batch_size_instances = max(0, rank_batch_size // self.task.max_sequence_length)
+
         log.info(
             f"Using per-rank batch size of {rank_batch_size_instances} instances "
             f"for downstream eval task '{task}' with max sequence length {self.task.max_sequence_length:,d} tokens"
@@ -215,9 +217,12 @@ def update_metrics(
         self.metric.update(batch, logits)
 
     def compute_metrics(self) -> Dict[str, torch.Tensor]:
-        value = self.metric.compute()
-        label = f"{self.label} ({self.metric_type_to_label[self.task.metric_type]})"
-        return {label: value}
+        metric_type_to_value = self.metric.compute()
+        outputs = {}
+        for metric_type, value in metric_type_to_value.items():
+            key = f"{self.label} ({self.metric_type_to_label[metric_type]})"
+            outputs[key] = value.item()
+        return outputs
 
     def reset_metrics(self) -> None:
         self.metric.reset()
@@ -227,7 +232,7 @@ def reset_metrics(self) -> None:
 class DownstreamEvaluatorCallbackConfig(CallbackConfig):
     tasks: List[str]
     tokenizer: TokenizerConfig
-    eval_batch_size: Optional[int] = None
+    eval_batch_size: Optional[int] = None  # NOTE: this counts in number of tokens
     eval_interval: int = 1000
     eval_duration: Duration = field(default_factory=lambda: Duration.epochs(1))
     log_interval: int = 5

src/scripts/train/OLMo2-1B.py

@@ -9,6 +9,9 @@
 from olmo_core.optim import AdamWConfig, OptimGroupOverride
 from olmo_core.train import TrainerConfig
 from olmo_core.train.callbacks import CheckpointerCallback, CometCallback, WandBCallback
+from olmo_core.train.callbacks.evaluator_callback import (
+    DownstreamEvaluatorCallbackConfig,
+)
 
 
 def build_model_config(common: CommonComponents) -> TransformerConfig:
@@ -73,6 +76,48 @@ def build_trainer_config(common: CommonComponents) -> TrainerConfig:
                 cancel_check_interval=10,
             ),
         )
+        .with_callback(
+            "downstream",
+            DownstreamEvaluatorCallbackConfig(
+                tasks=[
+                    "arc_challenge_val_rc_5shot",
+                    "arc_challenge_val_mc_5shot",
+                    "arc_challenge_test_rc_5shot",
+                    "arc_challenge_test_mc_5shot",
+                    "arc_easy_val_rc_5shot",
+                    "arc_easy_val_mc_5shot",
+                    "arc_easy_test_rc_5shot",
+                    "arc_easy_test_mc_5shot",
+                    "boolq_val_rc_5shot",
+                    "boolq_val_mc_5shot",
+                    "csqa_val_rc_5shot",
+                    "csqa_val_mc_5shot",
+                    "hellaswag_val_rc_5shot",
+                    "hellaswag_val_mc_5shot",
+                    "openbookqa_val_rc_5shot",
+                    "openbookqa_val_mc_5shot",
+                    "openbookqa_test_rc_5shot",
+                    "openbookqa_test_mc_5shot",
+                    "piqa_val_rc_5shot",
+                    "piqa_val_mc_5shot",
+                    "socialiqa_val_rc_5shot",
+                    "socialiqa_val_mc_5shot",
+                    "winogrande_val_rc_5shot",
+                    "winogrande_val_mc_5shot",
+                    "mmlu_stem_val_rc_5shot",
+                    "mmlu_stem_val_mc_5shot",
+                    "mmlu_humanities_val_rc_5shot",
+                    "mmlu_humanities_val_mc_5shot",
+                    "mmlu_social_sciences_val_rc_5shot",
+                    "mmlu_social_sciences_val_mc_5shot",
+                    "mmlu_other_val_rc_5shot",
+                    "mmlu_other_val_mc_5shot",
+                ],
+                tokenizer=common.tokenizer,
+                eval_batch_size=16 * 4096,
+                eval_interval=1,
+            ),
+        )
     )