Upload models/src/training/train_utils.py with huggingface_hub

models/src/training/train_utils.py

@@ -0,0 +1,69 @@
+import transformers
+import torch
+import logging
+
+
+def maybe_zero_3(param, ignore_status=False, name=None):
+    from deepspeed import zero
+    from deepspeed.runtime.zero.partition_parameters import ZeroParamStatus
+    if hasattr(param, "ds_id"):
+        if param.ds_status == ZeroParamStatus.NOT_AVAILABLE:
+            if not ignore_status:
+                logging.warning(f"{name}: param.ds_status != ZeroParamStatus.NOT_AVAILABLE: {param.ds_status}")
+        with zero.GatheredParameters([param]):
+            param = param.data.detach().cpu().clone()
+    else:
+        param = param.detach().cpu().clone()
+    return param
+
+# Borrowed from peft.utils.get_peft_model_state_dict
+def get_peft_state_maybe_zero_3(named_params, bias):
+    if bias == "none":
+        to_return = {k: t for k, t in named_params if "lora_" in k}
+    elif bias == "all":
+        to_return = {k: t for k, t in named_params if "lora_" in k or "bias" in k}
+    elif bias == "lora_only":
+        to_return = {}
+        maybe_lora_bias = {}
+        lora_bias_names = set()
+        for k, t in named_params:
+            if "lora_" in k:
+                to_return[k] = t
+                bias_name = k.split("lora_")[0] + "bias"
+                lora_bias_names.add(bias_name)
+            elif "bias" in k:
+                maybe_lora_bias[k] = t
+        for k, t in maybe_lora_bias:
+            if bias_name in lora_bias_names:
+                to_return[bias_name] = t
+    else:
+        raise NotImplementedError
+    to_return = {k: maybe_zero_3(v, ignore_status=True) for k, v in to_return.items()}
+    return to_return
+
+
+def get_peft_state_non_lora_maybe_zero_3(named_params, require_grad_only=True):
+    to_return = {k: t for k, t in named_params if "lora_" not in k}
+    if require_grad_only:
+        to_return = {k: t for k, t in to_return.items() if t.requires_grad}
+    to_return = {k: maybe_zero_3(v, ignore_status=True).cpu() for k, v in to_return.items()}
+    return to_return
+
+def safe_save_model_for_hf_trainer(trainer: transformers.Trainer,
+                                   output_dir: str):
+    """Collects the state dict and dump to disk."""
+
+    if trainer.deepspeed:
+        torch.cuda.synchronize()
+        trainer.save_model(output_dir)
+        return
+
+    state_dict = trainer.model.state_dict()
+    if trainer.args.should_save:
+        cpu_state_dict = {
+            key: value.cpu()
+            for key, value in state_dict.items()
+        }
+        del state_dict
+        trainer._save(output_dir, state_dict=cpu_state_dict)  # noqa
+        trainer.model.config.save_pretrained(output_dir)