lemonade-sdk 8.0.2__py3-none-any.whl → 8.0.4__py3-none-any.whl

{lemonade_sdk-8.0.2.dist-info → lemonade_sdk-8.0.4.dist-info}/RECORD

@@ -1,10 +1,10 @@
 lemonade/__init__.py,sha256=W1Qk7r0rnQqFhPNHp6BIBT_q-OH3s-8Q_POoVfAmKW0,117
 lemonade/api.py,sha256=X7DxBgsOl5L_z6uTkwoJWf8x0rjXWS2JoeEqmo9bMfc,3873
 lemonade/cache.py,sha256=djr2qgyUUAWlQv8FehU9qlNtCwK0IZqo82hcBDyZ3-A,2850
-lemonade/cli.py,sha256=XzptHh6LTl5OdGRnxiLykQ8QBl2rQmhWH5w0KPJVyY4,4359
+lemonade/cli.py,sha256=9Pcs3PcrWC2F8_pcBaz09xHUICIJTvpemBdPGyXkjIk,4395
 lemonade/sequence.py,sha256=KSH7BPsiyDKsOsg_ziQKEGsDwMmuO_YbgPRBxkZd0pw,13267
 lemonade/state.py,sha256=sdSezla7Cd7KYL90xY3p9kcNV4ndSyN6UvNLOr3vBMA,5261
-lemonade/version.py,sha256=hTIZ_8cc-ggqcFeOYQQKOHudFQCQNQlM4ZltuYIIjD4,22
+lemonade/version.py,sha256=8H4GfArMIlRTCgSsTERRXsD3PA6Y67z17oTQOJnuUME,22
 lemonade/common/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 lemonade/common/build.py,sha256=zTb0m1-kuUx6zw5QHp2SNnVuN6jOTMQ2FCdj9iH374U,6140
 lemonade/common/cli_helpers.py,sha256=hjBfXrTtFl8gmCFlL-ksviXR0mOcdPtTWVNKoEp3PG4,4993
@@ -17,16 +17,16 @@ lemonade/common/system_info.py,sha256=qOwteG_mBo-ImilbiK7Gq37sWIE9ugF0dbWcj9zLD4
 lemonade/common/test_helpers.py,sha256=Gwk-pa_6xYAo2oro-2EJNfuouAfw8k_brCbcMC-E-r0,758
 lemonade/profilers/__init__.py,sha256=JKVonvJ4XZ9_6sKXPWsiMLQCNyzQOxhQw5BEHR1qOfU,31
 lemonade/profilers/memory_tracker.py,sha256=1iuKt0FmNVYLDnOc-oZM8dX9TUksvoxO0m2EoYWjhYQ,9367
-lemonade/profilers/profiler.py,sha256=y_iMGr1ToQ6rcwcIcXck4ajapisLXCfHggiV-IpPF98,1666
+lemonade/profilers/profiler.py,sha256=Y5FSbc386bMlTVbqCuya9pYrso5aTthxahR1V_ZKQ9E,1902
 lemonade/tools/__init__.py,sha256=_6xRc-FHxmujoLjLjWtpYrWYEXtCSneSy-5ya01kyPk,53
 lemonade/tools/accuracy.py,sha256=9HCmczDngkBUuUrt49d2CkRo4J0qyWoFYs5cj20bGkg,11714
 lemonade/tools/adapter.py,sha256=HG54iMd6HDPZ4vnQIl7codq3HzffWbcHSIs_jVbNbhU,2958
 lemonade/tools/bench.py,sha256=aN5LMA_EH6-ZhAH3Gf26JYL7s0eKpUd3j-bReRhzvEY,10016
-lemonade/tools/humaneval.py,sha256=9lzsOaCSECf8LzqkQLFNwy1doAiZtK5gRN-RbZH7GLI,9532
+lemonade/tools/humaneval.py,sha256=JbxuoOzvR4iyxZv4R6MI7a3gUt5ef_Jj6Ie-9VP2wzY,9531
 lemonade/tools/management_tools.py,sha256=RO-lU-hjZhrP9KD9qcLI7MrLu-Rxnkrxzn45qqwKInE,8554
-lemonade/tools/mmlu.py,sha256=aEp9nMKTX5yaSaVZ15YmXbWE0YugjeAacnqjMZ13hHM,11072
-lemonade/tools/perplexity.py,sha256=xHl4cTBpJOCNcVxXhMv6eMp8fgUQmFM0G8DeRnx_rUk,5631
-lemonade/tools/prompt.py,sha256=AT3p5rCGHEs9ozeGxwWl07iKF-mgLxFOkYLjU2btFHs,8638
+lemonade/tools/mmlu.py,sha256=c2QaIMDzjqxCvgHlMXmy_dP1sAFkwkDxL7RO2nogI6s,11071
+lemonade/tools/perplexity.py,sha256=eiaTZ3yhqF2pfwOffVbKKJLwjSri7Im2pC-tBJr7LLU,5638
+lemonade/tools/prompt.py,sha256=cy6McZeLgk26xG1dJEY-cYnY2x8FUdyOOSG86WfBKCg,9348
 lemonade/tools/tool.py,sha256=UsxVYukfm_iM3BfeGYPZxQlTK5UfDfDOl3RIyLr8A1Y,13256
 lemonade/tools/huggingface/bench.py,sha256=-mTfldCtquL4mspq8ykVwDc9Mut5Ecv_jHJnSb0CYGE,6734
 lemonade/tools/huggingface/load.py,sha256=KsSGOBBD-tNEIfYC8mCWV_jpnkjHMhN3juVmC1Ln4uQ,7745
@@ -35,36 +35,36 @@ lemonade/tools/llamacpp/bench.py,sha256=A1X8ULQMxPVsff-AdiUsbWQUKpx7U7nFRNHFJRPd
 lemonade/tools/llamacpp/load.py,sha256=o3vVlefdxmdkHnuvFR3TOxiJkpNAuNFcs9Whfp24jpg,9236
 lemonade/tools/oga/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 lemonade/tools/oga/bench.py,sha256=T3c40NevM3NA7CT98B6vBj1nXfdITDqpfMHYSjhjwpA,5061
-lemonade/tools/oga/load.py,sha256=7Sdf6PFPrqbadPabyJb_uPRUIP09qj21ZYdXz47MqsE,28570
+lemonade/tools/oga/load.py,sha256=XSznW8lOX_KafSq5J5mIBJzj8YJEBpK0RFGcTE1wnE8,28317
 lemonade/tools/oga/utils.py,sha256=p7faMNfT-rLURC9t_s1S_STQRzzLADqbngUliTOOXeQ,16144
 lemonade/tools/quark/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lemonade/tools/quark/quark_load.py,sha256=tNy-G9yEJ5cTsxw9LmGUYmmdlEzMo_iy-KSIc2YVz6U,5581
-lemonade/tools/quark/quark_quantize.py,sha256=LZrcbLf9oIw7FW2ccP_qkCP32jxmz5YnNEaoY6rsAuY,16583
+lemonade/tools/quark/quark_load.py,sha256=FJ4LJKTToZbHHWVEOBLadae1a3jCnnY4KvXySHbkJMA,5589
+lemonade/tools/quark/quark_quantize.py,sha256=hwoaXhpBIORvJ16MvewphPkaDEQn3BAgXq5o82Gc-_s,16599
 lemonade/tools/report/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 lemonade/tools/report/llm_report.py,sha256=bVHhwCINA-Ok2EdSwAsLubsc83N3KWOVuwTguw7jDcE,6676
-lemonade/tools/report/table.py,sha256=VkTv5Vd0HOXudEthCBnFMrWK73Dm2AQP2_B83vEKBzI,25129
+lemonade/tools/report/table.py,sha256=wJFzKtlmGQH0RQ5O9nevtpMe_-zQ-8zNOndINQuzsjM,27793
 lemonade/tools/server/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lemonade/tools/server/llamacpp.py,sha256=aDVjjkU2Z2PN25Uuy-lk6ByKPR8kg5r2X-YsVSs4vi8,15624
-lemonade/tools/server/serve.py,sha256=3_jBpi6THnnAmtKOxvPlOkIhSTTmrlZE3fr2Dpto-Q4,52794
+lemonade/tools/server/llamacpp.py,sha256=e1MYKSJBu-jlOE5GQSBsC9CUPAeqw5wXXxoxBKA5zb8,20038
+lemonade/tools/server/serve.py,sha256=ORffC4bcBJ-L5-JbmZX91X3yHt1JWxZcIjrZuu9x8TQ,56165
 lemonade/tools/server/tool_calls.py,sha256=xrAlQwKG-nv2xLlf8f9CDSaUbyMn8ZtHkds9iZLG9K8,5230
-lemonade/tools/server/tray.py,sha256=OI2uCncs8UgnYFLCKHHXq06RETO2RFEcn4xLzMq-q_c,16675
+lemonade/tools/server/tray.py,sha256=4Kf3x8YfRaItPW7lxlEwerD7c5Q2snzcNk3ZrEoae58,17259
 lemonade/tools/server/webapp.py,sha256=8Das5yXOaSBLZmSZ_eddJajQFxBhvl5D6GI_hHlGbE0,1040
 lemonade/tools/server/static/favicon.ico,sha256=hMmP9qGJNeZ0mFS86JIqPbZstXMZn0Z76_HfHQpREAU,126745
-lemonade/tools/server/static/styles.css,sha256=u-SzZ-vh5qEFMDSKLHJ7MsQwvwpJLB_DdJxocf06Sro,16880
-lemonade/tools/server/static/webapp.html,sha256=im7YQkwvbuqrbO-sLhStVqtA6B7HKAn2azZka1KoeJQ,21260
+lemonade/tools/server/static/styles.css,sha256=x-pf7xts0te9JWAafcNFqzE7r1fl6n_H362Eiz49ixI,24722
+lemonade/tools/server/static/webapp.html,sha256=AS61ZBDnZkIUpT-iZFlTnWpkp6Yeozs4obzauX4crlU,35004
 lemonade/tools/server/utils/port.py,sha256=XnIg2qS73QRrsJn6LgHcrJPmku30Tv6vsYcBVMj82K4,2186
 lemonade/tools/server/utils/system_tray.py,sha256=b9lvNv9chJKQxvmH7qzAuUe6H9HsLu7pdHFqGlAJaL0,12654
 lemonade/tools/server/utils/thread.py,sha256=pK9K_6DNWoQ78NArkAX3Ym2WsxLnCs9sKTk6TitlYnI,2804
 lemonade_install/__init__.py,sha256=26zohKg2jgr_5y7tObduWMYQg8zCTWMZHL8lfi2zZVQ,40
 lemonade_install/install.py,sha256=DJWR36QSjZtvEwRjYPNSjhYgoxLjI_6OPrCMZjL0ChY,28263
-lemonade_sdk-8.0.2.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-lemonade_sdk-8.0.2.dist-info/licenses/NOTICE.md,sha256=B8lEqi4QE41J9ljz4Riv2JgHD1v8GCZE6nNBHO3KIA0,2135
+lemonade_sdk-8.0.4.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+lemonade_sdk-8.0.4.dist-info/licenses/NOTICE.md,sha256=B8lEqi4QE41J9ljz4Riv2JgHD1v8GCZE6nNBHO3KIA0,2135
 lemonade_server/cli.py,sha256=z6ojwFaOIz0hbUbVtZWMLP4YDpkcVOmqwmdm55dhKA4,11980
-lemonade_server/model_manager.py,sha256=HqbahDMRv1x8jyQj4pa1rXanlPmcCykt8tlI6WfaxjE,13023
-lemonade_server/pydantic_models.py,sha256=nsbpHqAkd6nkz5QT16u9xMZbCXqccGiy5O0fWecOM88,2338
-lemonade_server/server_models.json,sha256=wTK_H9XDHLxqMWQJqbBsJwm50PhOR4gURyVj9Jm35PQ,6992
-lemonade_sdk-8.0.2.dist-info/METADATA,sha256=hS5Xn5Pjq0RbdLlhedz3HQMCvkRrMWFoAI0Mao4cHwg,8225
-lemonade_sdk-8.0.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-lemonade_sdk-8.0.2.dist-info/entry_points.txt,sha256=gJppn0ETtXXR6ceKWEIRdk42kMC7ps59EmU3NCPyPUk,144
-lemonade_sdk-8.0.2.dist-info/top_level.txt,sha256=10ap5GNiPhalO4V50LRoxA1FqRT9g3Xkia6BITu880k,42
-lemonade_sdk-8.0.2.dist-info/RECORD,,
+lemonade_server/model_manager.py,sha256=0HqLR38uOu_hxRWVYQ_P6YmwaR-jkDuaAqGYo60X8C0,16702
+lemonade_server/pydantic_models.py,sha256=rp_FFhoTwg6jNmgol-kShwffnRDGbt7jTbIeELvgOIo,2876
+lemonade_server/server_models.json,sha256=Y-j9KAvHmfv77welC0rfRao4inLBce6AVySb-oy_uNE,7519
+lemonade_sdk-8.0.4.dist-info/METADATA,sha256=FqA9Jtgx1QE1EjLg_lxcfcAMI3j0cKpZxoe4GnaGLRA,7754
+lemonade_sdk-8.0.4.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+lemonade_sdk-8.0.4.dist-info/entry_points.txt,sha256=gJppn0ETtXXR6ceKWEIRdk42kMC7ps59EmU3NCPyPUk,144
+lemonade_sdk-8.0.4.dist-info/top_level.txt,sha256=10ap5GNiPhalO4V50LRoxA1FqRT9g3Xkia6BITu880k,42
+lemonade_sdk-8.0.4.dist-info/RECORD,,

lemonade_server/model_manager.py

@@ -54,6 +54,17 @@ class ModelManager:
                 for model_name, model_info in user_models.items()
             }
 
+            # Backwards compatibility for user models that were created before version 8.0.4
+            # "reasoning" was a boolean, but as of 8.0.4 it became a label
+            for _, model_info in user_models.items():
+                if "reasoning" in model_info:
+                    model_info["labels"] = (
+                        ["reasoning"]
+                        if not model_info["labels"]
+                        else model_info["labels"] + ["reasoning"]
+                    )
+                    del model_info["reasoning"]
+
             models.update(user_models)
 
         # Add the model name as a key in each entry, to make it easier
@@ -102,57 +113,131 @@ class ModelManager:
         """
         return self.filter_models_by_backend(self.downloaded_models)
 
+    def identify_gguf_models(
+        self, checkpoint: str, variant: str, mmproj: str
+    ) -> tuple[dict, list[str]]:
+        """
+        Identifies the GGUF model files in the repository that match the variant.
+        """
+
+        hint = """
+        The CHECKPOINT:VARIANT scheme is used to specify model files in Hugging Face repositories.
+
+        The VARIANT format can be one of several types:
+        1. Full filename: exact file to download
+        2. None/empty: gets the first .gguf file in the repository (excludes mmproj files)
+        3. Quantization variant: find a single file ending with the variant name (case insensitive)
+        4. Folder name: downloads all .gguf files in the folder that matches the variant name (case insensitive)
+        
+        Examples:
+        - "unsloth/Qwen3-8B-GGUF:qwen3.gguf" -> downloads "qwen3.gguf"
+        - "unsloth/Qwen3-30B-A3B-GGUF" -> downloads "Qwen3-30B-A3B-GGUF.gguf"
+        - "unsloth/Qwen3-8B-GGUF:Q4_1" -> downloads "Qwen3-8B-GGUF-Q4_1.gguf"
+        - "unsloth/Qwen3-30B-A3B-GGUF:Q4_0" -> downloads all files in "Q4_0/" folder
+        """
+
+        repo_files = huggingface_hub.list_repo_files(checkpoint)
+        sharded_files = []
+
+        # (case 1) If variant ends in .gguf, use it directly
+        if variant and variant.endswith(".gguf"):
+            variant_name = variant
+            if variant_name not in repo_files:
+                raise ValueError(
+                    f"File {variant} not found in Hugging Face repository {checkpoint}. {hint}"
+                )
+        # (case 2) If no variant is provided, get the first .gguf file in the repository
+        elif variant is None:
+            all_variants = [
+                f for f in repo_files if f.endswith(".gguf") and "mmproj" not in f
+            ]
+            if len(all_variants) == 0:
+                raise ValueError(
+                    f"No .gguf files found in Hugging Face repository {checkpoint}. {hint}"
+                )
+            variant_name = all_variants[0]
+        else:
+            # (case 3) Find a single file ending with the variant name (case insensitive)
+            end_with_variant = [
+                f
+                for f in repo_files
+                if f.lower().endswith(f"{variant}.gguf".lower())
+                and "mmproj" not in f.lower()
+            ]
+            if len(end_with_variant) == 1:
+                variant_name = end_with_variant[0]
+            elif len(end_with_variant) > 1:
+                raise ValueError(
+                    f"Multiple .gguf files found for variant {variant}, but only one is allowed. {hint}"
+                )
+            # (case 4) Check whether the variant corresponds to a folder with sharded files (case insensitive)
+            else:
+                sharded_files = [
+                    f
+                    for f in repo_files
+                    if f.endswith(".gguf")
+                    and f.lower().startswith(f"{variant}/".lower())
+                ]
+
+                if not sharded_files:
+                    raise ValueError(
+                        f"No .gguf files found for variant {variant}. {hint}"
+                    )
+
+                # Sort to ensure consistent ordering
+                sharded_files.sort()
+
+                # Use first file as primary (this is how llamacpp handles it)
+                variant_name = sharded_files[0]
+
+        core_files = {"variant": variant_name}
+
+        # If there is a mmproj file, add it to the patterns
+        if mmproj:
+            if mmproj not in repo_files:
+                raise ValueError(
+                    f"The provided mmproj file {mmproj} was not found in {checkpoint}."
+                )
+            core_files["mmproj"] = mmproj
+
+        return core_files, sharded_files
+
     def download_gguf(self, model_config: PullConfig) -> dict:
         """
         Downloads the GGUF file for the given model configuration.
+
+        For sharded models, if the variant points to a folder (e.g. Q4_0), all files in that folder
+        will be downloaded but only the first file will be returned for loading.
         """
 
-        # The variant parameter can be either:
-        # 1. A full GGUF filename (e.g. "model-Q4_0.gguf")
-        # 2. A quantization variant (e.g. "Q4_0")
-        # This code handles both cases by constructing the appropriate filename
+        # This code handles all cases by constructing the appropriate filename or pattern
         checkpoint, variant = self.parse_checkpoint(model_config.checkpoint)
-        hf_base_name = checkpoint.split("/")[-1].replace("-GGUF", "")
-        variant_name = (
-            variant if variant.endswith(".gguf") else f"{hf_base_name}-{variant}.gguf"
-        )
 
-        # If there is a mmproj file, add it to the patterns
-        expected_files = {"variant": variant_name}
-        if model_config.mmproj:
-            expected_files["mmproj"] = model_config.mmproj
+        # Identify the GGUF model files in the repository that match the variant
+        core_files, sharded_files = self.identify_gguf_models(
+            checkpoint, variant, model_config.mmproj
+        )
 
         # Download the files
         snapshot_folder = huggingface_hub.snapshot_download(
             repo_id=checkpoint,
-            allow_patterns=list(expected_files.values()),
+            allow_patterns=list(core_files.values()) + sharded_files,
         )
 
-        # Make sure we downloaded something
-        # If we didn't that can indicate that no patterns from allow_patterns match
-        # any files in the HF repo
-        if not os.path.exists(snapshot_folder):
-            raise ValueError(
-                "No patterns matched the variant parameter (CHECKPOINT:VARIANT). "
-                "Try again, providing the full filename of your target .gguf file as the variant."
-                " For example: Qwen/Qwen2.5-Coder-3B-Instruct-GGUF:"
-                "qwen2.5-coder-3b-instruct-q4_0.gguf"
-            )
-
-        # Ensure we downloaded all expected files while creating a dict of the downloaded files
-        snapshot_files = {}
-        for file in expected_files:
-            snapshot_files[file] = os.path.join(snapshot_folder, expected_files[file])
-            if expected_files[file].lower() not in [
-                name.lower() for name in os.listdir(snapshot_folder)
-            ]:
+        # Ensure we downloaded all expected files
+        for file in list(core_files.values()) + sharded_files:
+            expected_path = os.path.join(snapshot_folder, file)
+            if not os.path.exists(expected_path):
                 raise ValueError(
                     f"Hugging Face snapshot download for {model_config.checkpoint} "
-                    f"expected file {expected_files[file]} not found in {snapshot_folder}"
+                    f"expected file {file} not found at {expected_path}"
                 )
 
-        # Return a dict that points to the snapshot path of the downloaded GGUF files
-        return snapshot_files
+        # Return a dict of the full path of the core GGUF files
+        return {
+            file_name: os.path.join(snapshot_folder, file_path)
+            for file_name, file_path in core_files.items()
+        }
 
     def download_models(
         self,
@@ -194,9 +279,8 @@ class ModelManager:
                 new_user_model = {
                     "checkpoint": checkpoint,
                     "recipe": recipe,
-                    "reasoning": reasoning,
                     "suggested": True,
-                    "labels": ["custom"],
+                    "labels": ["custom"] + (["reasoning"] if reasoning else []),
                 }
 
                 if mmproj:
@@ -249,6 +333,9 @@ class ModelManager:
 
                 user_models[model_name] = new_user_model
 
+                # Ensure the cache directory exists before writing the file
+                os.makedirs(os.path.dirname(USER_MODELS_FILE), exist_ok=True)
+
                 with open(USER_MODELS_FILE, mode="w", encoding="utf-8") as file:
                     json.dump(user_models, fp=file)
 

lemonade_server/pydantic_models.py

@@ -1,4 +1,4 @@
-from typing import Optional
+from typing import Optional, Union, List, Any
 
 from pydantic import BaseModel
 
@@ -65,6 +65,30 @@ class ChatCompletionRequest(BaseModel):
     response_format: dict | None = None
 
 
+class EmbeddingsRequest(BaseModel):
+    """
+    Request model for embeddings API endpoint.
+
+    Generates embeddings for the provided input text or tokens.
+    """
+
+    input: Union[str, List]
+    model: Optional[str] = None
+    encoding_format: Optional[str] = "float"  # "float" or "base64"
+
+
+class RerankingRequest(BaseModel):
+    """
+    Request model for reranking API endpoint.
+
+    Reranks a list of documents based on their relevance to a query.
+    """
+
+    query: str
+    documents: List[str]
+    model: str
+
+
 class ResponsesRequest(BaseModel):
     """
     Request model for responses API endpoint.

lemonade_server/server_models.json

@@ -2,197 +2,177 @@
     "Qwen2.5-0.5B-Instruct-CPU": {
         "checkpoint": "amd/Qwen2.5-0.5B-Instruct-quantized_int4-float16-cpu-onnx",
         "recipe": "oga-cpu",
-        "reasoning": false,
         "suggested": true
     },
     "Llama-3.2-1B-Instruct-CPU": {
         "checkpoint": "amd/Llama-3.2-1B-Instruct-awq-uint4-float16-cpu-onnx",
         "recipe": "oga-cpu",
-        "reasoning": false,
         "suggested": false
     },
     "Llama-3.2-3B-Instruct-CPU": {
         "checkpoint": "amd/Llama-3.2-3B-Instruct-awq-uint4-float16-cpu-onnx",
         "recipe": "oga-cpu",
-        "reasoning": false,
         "suggested": false
     },
     "Phi-3-Mini-Instruct-CPU": {
         "checkpoint": "amd/Phi-3-mini-4k-instruct_int4_float16_onnx_cpu",
         "recipe": "oga-cpu",
-        "reasoning": false,
         "suggested": true
     },
     "Qwen-1.5-7B-Chat-CPU": {
         "checkpoint": "amd/Qwen1.5-7B-Chat_uint4_asym_g128_float16_onnx_cpu",
         "recipe": "oga-cpu",
-        "reasoning": false,
         "suggested": true
     },
     "DeepSeek-R1-Distill-Llama-8B-CPU": {
         "checkpoint": "amd/DeepSeek-R1-Distill-Llama-8B-awq-asym-uint4-g128-lmhead-onnx-cpu",
         "recipe": "oga-cpu",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "DeepSeek-R1-Distill-Qwen-7B-CPU": {
         "checkpoint": "amd/DeepSeek-R1-Distill-Llama-8B-awq-asym-uint4-g128-lmhead-onnx-cpu",
         "recipe": "oga-cpu",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Llama-3.2-1B-Instruct-Hybrid": {
         "checkpoint": "amd/Llama-3.2-1B-Instruct-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 3000,
         "suggested": true
     },
     "Llama-3.2-3B-Instruct-Hybrid": {
         "checkpoint": "amd/Llama-3.2-3B-Instruct-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 2000,
         "suggested": true
     },
     "Phi-3-Mini-Instruct-Hybrid": {
         "checkpoint": "amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 2000,
         "suggested": true
     },
     "Phi-3.5-Mini-Instruct-Hybrid": {
         "checkpoint": "amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "suggested": false
     },
     "Qwen-1.5-7B-Chat-Hybrid": {
         "checkpoint": "amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 3000,
         "suggested": true
     },
     "DeepSeek-R1-Distill-Llama-8B-Hybrid": {
         "checkpoint": "amd/DeepSeek-R1-Distill-Llama-8B-awq-asym-uint4-g128-lmhead-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": true,
         "max_prompt_length": 2000,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "DeepSeek-R1-Distill-Qwen-7B-Hybrid": {
         "checkpoint": "amd/DeepSeek-R1-Distill-Qwen-7B-awq-asym-uint4-g128-lmhead-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": true,
         "max_prompt_length": 2000,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Mistral-7B-v0.3-Instruct-Hybrid": {
         "checkpoint": "amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 2000,
         "suggested": true
     },
     "Llama-3.1-8B-Instruct-Hybrid": {
         "checkpoint": "amd/Llama-3.1-8B-Instruct-awq-asym-uint4-g128-lmhead-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 2000,
         "suggested": true
     },
     "Llama-xLAM-2-8b-fc-r-Hybrid": {
         "checkpoint": "amd/Llama-xLAM-2-8b-fc-r-awq-g128-int4-asym-bfp16-onnx-hybrid",
         "recipe": "oga-hybrid",
-        "reasoning": false,
         "max_prompt_length": 2000,
         "suggested": true
     },
     "Llama-3.2-1B-Instruct-DirectML": {
         "checkpoint": "amd/Llama-3.2-1B-Instruct-dml-int4-awq-block-128-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Llama-3.2-3B-Instruct-DirectML": {
         "checkpoint": "amd/Llama-3.2-3B-Instruct-dml-int4-awq-block-128-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Phi-3.5-Mini-Instruct-DirectML": {
         "checkpoint": "amd/phi3.5-mini-instruct-int4-awq-block-128-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Qwen-1.5-7B-Chat-DirectML": {
         "checkpoint": "amd/Qwen1.5-7B-Chat-dml-int4-awq-block-128-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Mistral-7B-v0.1-Instruct-DirectML": {
         "checkpoint": "amd/Mistral-7B-Instruct-v0.1-awq-g128-int4-onnx-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Llama-3-8B-Instruct-DirectML": {
         "checkpoint": "amd/llama3-8b-instruct-awq-g128-int4-onnx-directml",
         "recipe": "oga-igpu",
-        "reasoning": false,
         "suggested": false
     },
     "Qwen3-0.6B-GGUF": {
         "checkpoint": "unsloth/Qwen3-0.6B-GGUF:Q4_0",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Qwen3-1.7B-GGUF": {
         "checkpoint": "unsloth/Qwen3-1.7B-GGUF:Q4_0",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Qwen3-4B-GGUF": {
         "checkpoint": "unsloth/Qwen3-4B-GGUF:Q4_0",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Qwen3-8B-GGUF": {
         "checkpoint": "unsloth/Qwen3-8B-GGUF:Q4_1",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "DeepSeek-Qwen3-8B-GGUF": {
         "checkpoint": "unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_1",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Qwen3-14B-GGUF": {
         "checkpoint": "unsloth/Qwen3-14B-GGUF:Q4_0",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Qwen3-30B-A3B-GGUF": {
         "checkpoint": "unsloth/Qwen3-30B-A3B-GGUF:Q4_0",
         "recipe": "llamacpp",
-        "reasoning": true,
-        "suggested": true
+        "suggested": true,
+        "labels": ["reasoning"]
     },
     "Gemma-3-4b-it-GGUF": {
         "checkpoint": "ggml-org/gemma-3-4b-it-GGUF:Q4_K_M",
         "mmproj": "mmproj-model-f16.gguf",
         "recipe": "llamacpp",
-        "reasoning": false,
         "suggested": true,
         "labels": ["vision"]
     },
@@ -200,8 +180,38 @@
         "checkpoint": "ggml-org/Qwen2.5-VL-7B-Instruct-GGUF:Q4_K_M",
         "mmproj": "mmproj-Qwen2.5-VL-7B-Instruct-f16.gguf",
         "recipe": "llamacpp",
-        "reasoning": false,
         "suggested": true,
         "labels": ["vision"]
+    },
+    "Llama-4-Scout-17B-16E-Instruct-GGUF": {
+        "checkpoint": "unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF:Q4_K_S",
+        "mmproj": "mmproj-F16.gguf",
+        "recipe": "llamacpp",
+        "suggested": true,
+        "labels": ["vision"]
+    },
+    "nomic-embed-text-v1-GGUF": {
+        "checkpoint": "nomic-ai/nomic-embed-text-v1-GGUF:Q4_K_S",
+        "recipe": "llamacpp",
+        "suggested": true,
+        "labels": ["embeddings"]
+    },
+    "nomic-embed-text-v2-moe-GGUF": {
+        "checkpoint": "nomic-ai/nomic-embed-text-v2-moe-GGUF:Q8_0",
+        "recipe": "llamacpp",
+        "suggested": true,
+        "labels": ["embeddings"]
+    },
+    "bge-reranker-v2-m3-GGUF": {
+        "checkpoint": "pqnet/bge-reranker-v2-m3-Q8_0-GGUF",
+        "recipe": "llamacpp",
+        "suggested": true,
+        "labels": ["reranking"]
+    },
+    "jina-reranker-v1-tiny-en-GGUF": {
+        "checkpoint": "mradermacher/jina-reranker-v1-tiny-en-GGUF:Q8_0",
+        "recipe": "llamacpp",
+        "suggested": false,
+        "labels": ["reranking"]
     }
 }