Model: Fix assorted bugs

and ignore warnings in pytest

modeling/inference_model.py

@@ -552,9 +552,6 @@ class InferenceModel:
         assert isinstance(prompt_tokens, np.ndarray)
         assert len(prompt_tokens.shape) == 1
 
-        if utils.koboldai_vars.model == "ReadOnly":
-            raise NotImplementedError("No loaded model")
-
         time_start = time.time()
 
         with use_core_manipulations():

modeling/inference_models/generic_hf_torch.py

@@ -33,10 +33,11 @@ class GenericHFTorchInferenceModel(HFTorchInferenceModel):
         # if utils.koboldai_vars.model not in ["NeoCustom", "GPT2Custom"]:
         #     utils.koboldai_vars.custmodpth = utils.koboldai_vars.model
 
-        if utils.koboldai_vars.model == "NeoCustom":
-            utils.koboldai_vars.model = os.path.basename(
+        if self.model_name == "NeoCustom":
+            self.model_name = os.path.basename(
                 os.path.normpath(utils.koboldai_vars.custmodpth)
             )
+            utils.koboldai_vars.model = self.model_name
 
         # If we specify a model and it's in the root directory, we need to move
         # it to the models directory (legacy folder structure to new)
@@ -123,8 +124,8 @@ class GenericHFTorchInferenceModel(HFTorchInferenceModel):
                     return old_rebuild_tensor(storage, storage_offset, shape, stride)
 
                 torch._utils._rebuild_tensor = new_rebuild_tensor
-                self.model = self._get_model(utils.koboldai_vars.model, tf_kwargs)
-                self.tokenizer = self._get_tokenizer(utils.koboldai_vars.model)
+                self.model = self._get_model(self.model_name, tf_kwargs)
+                self.tokenizer = self._get_tokenizer(self.model_name)
                 torch._utils._rebuild_tensor = old_rebuild_tensor
 
                 if save_model:
@@ -153,7 +154,7 @@ class GenericHFTorchInferenceModel(HFTorchInferenceModel):
                         shutil.move(
                             os.path.realpath(
                                 huggingface_hub.hf_hub_download(
-                                    utils.koboldai_vars.model,
+                                    self.model_name,
                                     transformers.configuration_utils.CONFIG_NAME,
                                     revision=utils.koboldai_vars.revision,
                                     cache_dir="cache",
@@ -177,7 +178,7 @@ class GenericHFTorchInferenceModel(HFTorchInferenceModel):
                                     shutil.move(
                                         os.path.realpath(
                                             huggingface_hub.hf_hub_download(
-                                                utils.koboldai_vars.model,
+                                                self.model_name,
                                                 possible_weight_name,
                                                 revision=utils.koboldai_vars.revision,
                                                 cache_dir="cache",
@@ -214,7 +215,7 @@ class GenericHFTorchInferenceModel(HFTorchInferenceModel):
                                 shutil.move(
                                     os.path.realpath(
                                         huggingface_hub.hf_hub_download(
-                                            utils.koboldai_vars.model,
+                                            self.model_name,
                                             filename,
                                             revision=utils.koboldai_vars.revision,
                                             cache_dir="cache",

modeling/inference_models/hf.py

@@ -8,9 +8,13 @@ from modeling.inference_model import InferenceModel
 
 
 class HFInferenceModel(InferenceModel):
-    def __init__(self) -> None:
+    def __init__(self, model_name: str) -> None:
         super().__init__()
         self.model_config = None
+        self.model_name = model_name
+
+        self.model = None
+        self.tokenizer = None
 
     def get_local_model_path(
         self, legacy: bool = False, ignore_existance: bool = False
@@ -34,7 +38,7 @@ class HFInferenceModel(InferenceModel):
         # Get the model_type from the config or assume a model type if it isn't present
         try:
             self.model_config = AutoConfig.from_pretrained(
-                self.get_local_model_path() or utils.koboldai_vars.model,
+                self.get_local_model_path() or self.model_name,
                 revision=utils.koboldai_vars.revision,
                 cache_dir="cache",
             )

modeling/inference_models/hf_mtj.py

@@ -30,12 +30,8 @@ class HFMTJInferenceModel(HFInferenceModel):
         self,
         model_name: str,
     ) -> None:
-        super().__init__()
+        super().__init__(model_name)
 
-        self.model_name = model_name
-
-        self.model = None
-        self.tokenizer = None
         self.model_config = None
         self.capabilties = ModelCapabilities(
             embedding_manipulation=False,

modeling/inference_models/hf_torch.py

@@ -61,9 +61,7 @@ class HFTorchInferenceModel(HFInferenceModel):
         lazy_load: bool,
         low_mem: bool,
     ) -> None:
-        super().__init__()
-
-        self.model_name = model_name
+        super().__init__(model_name)
         self.lazy_load = lazy_load
         self.low_mem = low_mem
 
@@ -78,8 +76,6 @@ class HFTorchInferenceModel(HFInferenceModel):
             Stoppers.chat_mode_stopper,
         ]
 
-        self.model = None
-        self.tokenizer = None
         self.capabilties = ModelCapabilities(
             embedding_manipulation=True,
             post_token_hooks=True,

pytest.ini

@@ -1,3 +1,3 @@
 [pytest]
-addopts = --ignore=miniconda3 --ignore=runtime --html=unit_test_report.html --self-contained-html -vv
+addopts = --ignore=miniconda3 --ignore=runtime --html=unit_test_report.html --self-contained-html --disable-warnings -vv
 norecursedirs = .git