Sync contents

.sync_state

@@ -1,4 +1,4 @@
 {
-  "last_synced_sha": "896d014e55e38b513967cd3cb9672200a421ff9f",
+  "last_synced_sha": "e99a92629bc1b7df95b2ffc0b60a60c8eecddee1",
-  "last_sync_time": "2026-01-16T05:35:31.063947"
+  "last_sync_time": "2026-01-21T23:13:02.930293"
 }

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "tinker"
-version = "0.8.0"
+version = "0.8.1"
 description = "The official Python SDK for the tinker API"
 readme = "README.md"
 license = "Apache-2.0"

src/tinker/_compat.py

@@ -116,10 +116,12 @@ def get_model_fields(model: type[pydantic.BaseModel]) -> dict[str, FieldInfo]:
     return model.__fields__  # type: ignore
 
 
-def model_copy(model: _ModelT, *, deep: bool = False) -> _ModelT:
+def model_copy(
+    model: _ModelT, *, deep: bool = False, update: dict[str, Any] | None = None
+) -> _ModelT:
     if PYDANTIC_V2:
-        return model.model_copy(deep=deep)
+        return model.model_copy(deep=deep, update=update)  # type: ignore
-    return model.copy(deep=deep)  # type: ignore
+    return model.copy(deep=deep, update=update)  # type: ignore
 
 
 def model_json(model: pydantic.BaseModel, *, indent: int | None = None) -> str:

src/tinker/lib/public_interfaces/sampling_client.py

@@ -370,6 +370,8 @@ def _load_tokenizer_from_model_info(
     """
     from transformers.models.auto.tokenization_auto import AutoTokenizer
 
+    model_name = model_name.split(":")[0]
+
     # Use tokenizer_id if provided, otherwise fall back to heuristic logic
     kwargs = {}
     if tokenizer_id is None:

src/tinker/lib/public_interfaces/training_client.py

@@ -27,15 +27,12 @@ from ..sync_only import sync_only
 from .sampling_client import SamplingClient, _load_tokenizer_from_model_info
 
 try:
-    import torch  # type: ignore[import-not-found]
+    import torch
-
-    _HAVE_TORCH = True
 except ImportError:
-    _HAVE_TORCH = False
+    torch = None
 
 
 if TYPE_CHECKING:
-    import torch
     from transformers.tokenization_utils import PreTrainedTokenizer
 
     from ..internal_client_holder import InternalClientHolder
@@ -372,13 +369,13 @@ class TrainingClient(TelemetryProvider, QueueStateObserver):
         self, data: List[types.Datum], loss_fn: CustomLossFnV1
     ) -> APIFuture[types.ForwardBackwardOutput]:
         """Async version of forward_backward_custom."""
-        if not _HAVE_TORCH:
+        if torch is None:
             raise ImportError("PyTorch is not installed. Cannot run custom forward_backward.")
 
         # First do a forward pass and get logprobs
         forward_future = await self.forward_async(data, "cross_entropy")
         forward_result = await forward_future.result_async()
-        logprobs_list: List["torch.Tensor"] = []
+        logprobs_list = []
         for out in forward_result.loss_fn_outputs:
             logprob = torch.tensor(out["logprobs"].data).clone().detach().requires_grad_(True)
             logprobs_list.append(logprob)
@@ -423,6 +420,11 @@ class TrainingClient(TelemetryProvider, QueueStateObserver):
     def optim_step(self, adam_params: types.AdamParams) -> APIFuture[types.OptimStepResponse]:
         """Update model parameters using Adam optimizer.
 
+        The Adam optimizer used by tinker is identical
+        to [torch.optim.AdamW](https://docs.pytorch.org/docs/stable/generated/torch.optim.AdamW.html).
+        Note that unlike PyTorch, Tinker's default weight decay value is 0.0 (no weight decay).
+
+
         Args:
         - `adam_params`: Adam optimizer parameters (learning_rate, betas, eps, weight_decay)
 

src/tinker/resources/models.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import httpx
 
 from .._base_client import make_request_options
-from .._compat import model_dump
+from .._compat import model_copy, model_dump
 from .._resource import AsyncAPIResource
 from .._types import NOT_GIVEN, Body, Headers, NotGiven, Query
 from ..types.create_model_request import CreateModelRequest
@@ -104,12 +104,19 @@ class AsyncModelsResource(AsyncAPIResource):
         if max_retries is not NOT_GIVEN:
             options["max_retries"] = max_retries
 
-        return await self._post(
+        result = await self._post(
             "/api/v1/get_info",
             body=model_dump(request, exclude_unset=True, mode="json"),
             options=options,
             cast_to=GetInfoResponse,
         )
+        if result.model_data.tokenizer_id:
+            tokenizer_id = result.model_data.tokenizer_id.split(":")[0]
+            updated_model_data = model_copy(
+                result.model_data, update={"tokenizer_id": tokenizer_id}
+            )
+            result = model_copy(result, update={"model_data": updated_model_data})
+        return result
 
     async def unload(
         self,

src/tinker/types/optim_step_request.py

@@ -26,7 +26,7 @@ class AdamParams(StrictBase):
     """Weight decay for the optimizer. Uses decoupled weight decay."""
 
     grad_clip_norm: float = 0.0
-    """Gradient clip norm for the optimizer. 0.0 means no clipping."""
+    """Maximum global gradient norm. If the global gradient norm is greater than this value, it will be clipped to this value. 0.0 means no clipping."""
 
 
 class OptimStepRequest(StrictBase):