together 1.5.34__tar.gz → 1.5.35__tar.gz

{together-1.5.34 → together-1.5.35}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: together
-Version: 1.5.34
+Version: 1.5.35
 Summary: Python client for Together's Cloud Platform! Note: SDK 2.0 is now available at https://github.com/togethercomputer/together-py
 License: Apache-2.0
 License-File: LICENSE

{together-1.5.34 → together-1.5.35}/pyproject.toml

@@ -12,7 +12,7 @@ build-backend = "poetry.masonry.api"
 
 [tool.poetry]
 name = "together"
-version = "1.5.34"
+version = "1.5.35"
 authors = ["Together AI <support@together.ai>"]
 description = "Python client for Together's Cloud Platform! Note: SDK 2.0 is now available at https://github.com/togethercomputer/together-py"
 readme = "README.md"

{together-1.5.34 → together-1.5.35}/src/together/cli/api/endpoints.py

@@ -98,7 +98,7 @@ def endpoints(ctx: click.Context) -> None:
 )
 @click.option(
     "--gpu",
-    type=click.Choice(["h100", "a100", "l40", "l40s", "rtx-6000"]),
+    type=click.Choice(["b200", "h200", "h100", "a100", "l40", "l40s", "rtx-6000"]),
     required=True,
     help="GPU type to use for inference",
 )
@@ -161,6 +161,8 @@ def create(
     """Create a new dedicated inference endpoint."""
     # Map GPU types to their full hardware ID names
     gpu_map = {
+        "b200": "nvidia_b200_180gb_sxm",
+        "h200": "nvidia_h200_140gb_sxm",
         "h100": "nvidia_h100_80gb_sxm",
         "a100": "nvidia_a100_80gb_pcie" if gpu_count == 1 else "nvidia_a100_80gb_sxm",
         "l40": "nvidia_l40",

{together-1.5.34 → together-1.5.35}/src/together/cli/api/utils.py

@@ -103,13 +103,13 @@ def generate_progress_bar(
     progress = "Progress: [bold red]unavailable[/bold red]"
     if finetune_job.status in COMPLETED_STATUSES:
         progress = "Progress: [bold green]completed[/bold green]"
-    elif finetune_job.updated_at is not None:
+    elif finetune_job.started_at is not None:
         # Replace 'Z' with '+00:00' for Python 3.10 compatibility
-        updated_at_str = finetune_job.updated_at.replace("Z", "+00:00")
-        update_at = datetime.fromisoformat(updated_at_str).astimezone()
+        started_at_str = finetune_job.started_at.replace("Z", "+00:00")
+        started_at = datetime.fromisoformat(started_at_str).astimezone()
 
         if finetune_job.progress is not None:
-            if current_time < update_at:
+            if current_time < started_at:
                 return progress
 
             if not finetune_job.progress.estimate_available:
@@ -118,7 +118,7 @@ def generate_progress_bar(
             if finetune_job.progress.seconds_remaining <= 0:
                 return progress
 
-            elapsed_time = (current_time - update_at).total_seconds()
+            elapsed_time = (current_time - started_at).total_seconds()
             ratio_filled = min(
                 elapsed_time / finetune_job.progress.seconds_remaining, 1.0
             )

{together-1.5.34 → together-1.5.35}/src/together/resources/endpoints.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import warnings
 from typing import Dict, List, Literal, Optional, Union
 
 from together.abstract import api_requestor
@@ -96,6 +97,12 @@ class Endpoints:
         Returns:
             DedicatedEndpoint: Object containing endpoint information
         """
+        if disable_prompt_cache:
+            warnings.warn(
+                "The 'disable_prompt_cache' parameter (CLI flag: '--no-prompt-cache') is deprecated and will be removed in a future version.",
+                stacklevel=2,
+            )
+
         requestor = api_requestor.APIRequestor(
             client=self._client,
         )
@@ -389,6 +396,12 @@ class AsyncEndpoints:
         Returns:
             DedicatedEndpoint: Object containing endpoint information
         """
+        if disable_prompt_cache:
+            warnings.warn(
+                "The 'disable_prompt_cache' parameter (CLI flag: '--no-prompt-cache') is deprecated and will be removed in a future version.",
+                stacklevel=2,
+            )
+
         requestor = api_requestor.APIRequestor(
             client=self._client,
         )

{together-1.5.34 → together-1.5.35}/src/together/types/finetune.py

@@ -286,6 +286,7 @@ class FinetuneResponse(BaseModel):
     # created/updated datetime stamps
     created_at: str | None = None
     updated_at: str | None = None
+    started_at: str | None = None
     # job status
     status: FinetuneJobStatus | None = None
     # job id