diff --git a/common/sampling.py b/common/sampling.py
index 2305316..6c22bfd 100644
--- a/common/sampling.py
+++ b/common/sampling.py
@@ -23,90 +23,90 @@ class BaseSamplerRequest(BaseModel):
         examples=[512],
     )
 
-    stop: Union[str, List[str]] = Field(
+    stop: Optional[Union[str, List[str]]] = Field(
         default_factory=lambda: get_default_sampler_value("stop", [])
     )
 
-    token_healing: bool = Field(
+    token_healing: Optional[bool] = Field(
         default_factory=lambda: get_default_sampler_value("token_healing", False)
     )
 
-    temperature: float = Field(
+    temperature: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("temperature", 1.0),
         examples=[1.0],
     )
 
-    temperature_last: bool = Field(
+    temperature_last: Optional[bool] = Field(
         default_factory=lambda: get_default_sampler_value("temperature_last", False)
     )
 
-    smoothing_factor: float = Field(
+    smoothing_factor: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("smoothing_factor", 0.0),
     )
 
-    top_k: int = Field(
+    top_k: Optional[int] = Field(
         default_factory=lambda: get_default_sampler_value("top_k", 0),
     )
 
-    top_p: float = Field(
+    top_p: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("top_p", 1.0),
         examples=[1.0],
     )
 
-    top_a: float = Field(
+    top_a: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("top_a", 0.0)
     )
 
-    min_p: float = Field(
+    min_p: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("min_p", 0.0)
     )
 
-    tfs: float = Field(
+    tfs: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("tfs", 1.0),
         examples=[1.0],
     )
 
-    frequency_penalty: float = Field(
+    frequency_penalty: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("frequency_penalty", 0.0)
     )
 
-    presence_penalty: float = Field(
+    presence_penalty: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("presence_penalty", 0.0)
     )
 
-    repetition_penalty: float = Field(
+    repetition_penalty: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("repetition_penalty", 1.0),
         examples=[1.0],
     )
 
-    repetition_decay: int = Field(
+    repetition_decay: Optional[int] = Field(
         default_factory=lambda: get_default_sampler_value("repetition_decay", 0)
     )
 
-    mirostat_mode: int = Field(
+    mirostat_mode: Optional[int] = Field(
         default_factory=lambda: get_default_sampler_value("mirostat_mode", 0)
     )
 
-    mirostat_tau: float = Field(
+    mirostat_tau: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("mirostat_tau", 1.5),
         examples=[1.5],
     )
 
-    mirostat_eta: float = Field(
+    mirostat_eta: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("mirostat_eta", 0.3),
         examples=[0.3],
     )
 
-    add_bos_token: bool = Field(
+    add_bos_token: Optional[bool] = Field(
         default_factory=lambda: get_default_sampler_value("add_bos_token", True)
     )
 
-    ban_eos_token: bool = Field(
+    ban_eos_token: Optional[bool] = Field(
         default_factory=lambda: get_default_sampler_value("ban_eos_token", False),
         examples=[False],
     )
 
-    skip_special_tokens: bool = Field(
+    skip_special_tokens: Optional[bool] = Field(
         default_factory=lambda: get_default_sampler_value("ban_eos_token", True),
         examples=[True],
     )
@@ -133,14 +133,14 @@ class BaseSamplerRequest(BaseModel):
     )
 
     # Aliased variables
-    typical: float = Field(
+    typical: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("typical", 1.0),
         validation_alias=AliasChoices("typical", "typical_p"),
         description="Aliases: typical_p",
         examples=[1.0],
     )
 
-    penalty_range: int = Field(
+    penalty_range: Optional[int] = Field(
         default_factory=lambda: get_default_sampler_value("penalty_range", -1),
         validation_alias=AliasChoices(
             "penalty_range",
@@ -150,34 +150,34 @@ class BaseSamplerRequest(BaseModel):
         description="Aliases: repetition_range, repetition_penalty_range",
     )
 
-    cfg_scale: float = Field(
+    cfg_scale: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("cfg_scale", 1.0),
         validation_alias=AliasChoices("cfg_scale", "guidance_scale"),
         description="Aliases: guidance_scale",
         examples=[1.0],
     )
 
-    max_temp: float = Field(
+    max_temp: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("max_temp", 1.0),
         validation_alias=AliasChoices("max_temp", "dynatemp_high"),
         description="Aliases: dynatemp_high",
         examples=[1.0],
     )
 
-    min_temp: float = Field(
+    min_temp: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("min_temp", 1.0),
         validation_alias=AliasChoices("min_temp", "dynatemp_low"),
         description="Aliases: dynatemp_low",
         examples=[1.0],
     )
 
-    temp_exponent: float = Field(
+    temp_exponent: Optional[float] = Field(
         default_factory=lambda: get_default_sampler_value("temp_exponent", 1.0),
         validation_alias=AliasChoices("temp_exponent", "dynatemp_exponent"),
         examples=[1.0],
     )
 
-    banned_tokens: Union[List[int], str] = Field(
+    banned_tokens: Optional[Union[List[int], str]] = Field(
         default_factory=lambda: get_default_sampler_value("banned_tokens", []),
         validation_alias=AliasChoices("banned_tokens", "custom_token_bans"),
         description="Aliases: custom_token_bans",
diff --git a/endpoints/OAI/types/chat_completion.py b/endpoints/OAI/types/chat_completion.py
index c264876..92265a7 100644
--- a/endpoints/OAI/types/chat_completion.py
+++ b/endpoints/OAI/types/chat_completion.py
@@ -43,8 +43,8 @@ class ChatCompletionRequest(CommonCompletionRequest):
     # Take in a string as well even though it's not part of the OAI spec
     messages: Union[str, List[Dict[str, str]]]
     prompt_template: Optional[str] = None
-    add_generation_prompt: bool = True
-    template_vars: dict = {}
+    add_generation_prompt: Optional[bool] = True
+    template_vars: Optional[dict] = {}
     response_prefix: Optional[str] = None
 
 
diff --git a/endpoints/OAI/types/common.py b/endpoints/OAI/types/common.py
index 0c3fe8f..b9aac68 100644
--- a/endpoints/OAI/types/common.py
+++ b/endpoints/OAI/types/common.py
@@ -26,8 +26,8 @@ class CommonCompletionRequest(BaseSamplerRequest):
     model: Optional[str] = None
 
     # Generation info (remainder is in BaseSamplerRequest superclass)
-    stream: bool = False
-    logprobs: int = 0
+    stream: Optional[bool] = False
+    logprobs: Optional[int] = 0
     response_format: Optional[CompletionResponseFormat] = Field(
         default_factory=CompletionResponseFormat
     )
@@ -36,10 +36,12 @@ class CommonCompletionRequest(BaseSamplerRequest):
     best_of: Optional[int] = Field(
         description="Not parsed. Only used for OAI compliance.", default=None
     )
-    echo: bool = Field(
+    echo: Optional[bool] = Field(
         description="Not parsed. Only used for OAI compliance.", default=False
     )
-    n: int = Field(description="Not parsed. Only used for OAI compliance.", default=1)
+    n: Optional[int] = Field(
+        description="Not parsed. Only used for OAI compliance.", default=1
+    )
     suffix: Optional[str] = Field(
         description="Not parsed. Only used for OAI compliance.", default=None
     )
diff --git a/endpoints/OAI/types/lora.py b/endpoints/OAI/types/lora.py
index b9e29bc..8435a8a 100644
--- a/endpoints/OAI/types/lora.py
+++ b/endpoints/OAI/types/lora.py
@@ -26,7 +26,7 @@ class LoraLoadInfo(BaseModel):
     """Represents a single Lora load info."""
 
     name: str
-    scaling: float = 1.0
+    scaling: Optional[float] = 1.0
 
 
 class LoraLoadRequest(BaseModel):
diff --git a/endpoints/OAI/types/model.py b/endpoints/OAI/types/model.py
index c769960..22895be 100644
--- a/endpoints/OAI/types/model.py
+++ b/endpoints/OAI/types/model.py
@@ -13,10 +13,10 @@ class ModelCardParameters(BaseModel):
     # Safe to do this since it's guaranteed to fetch a max seq len
     # from model_container
     max_seq_len: Optional[int] = None
-    rope_scale: float = 1.0
-    rope_alpha: float = 1.0
-    cache_mode: str = "FP16"
-    chunk_size: int = 2048
+    rope_scale: Optional[float] = 1.0
+    rope_alpha: Optional[float] = 1.0
+    cache_mode: Optional[str] = "FP16"
+    chunk_size: Optional[int] = 2048
     prompt_template: Optional[str] = None
     num_experts_per_token: Optional[int] = None
     use_cfg: Optional[bool] = None
@@ -47,7 +47,7 @@ class DraftModelLoadRequest(BaseModel):
     """Represents a draft model load request."""
 
     draft_model_name: str
-    draft_rope_scale: float = 1.0
+    draft_rope_scale: Optional[float] = 1.0
     draft_rope_alpha: Optional[float] = Field(
         description="Automatically calculated if not present",
         default=None,
@@ -73,9 +73,11 @@ class ModelLoadRequest(BaseModel):
         default=None,
         examples=[4096],
     )
-    gpu_split_auto: bool = True
-    autosplit_reserve: List[float] = [96]
-    gpu_split: List[float] = Field(default_factory=list, examples=[[24.0, 20.0]])
+    gpu_split_auto: Optional[bool] = True
+    autosplit_reserve: Optional[List[float]] = [96]
+    gpu_split: Optional[List[float]] = Field(
+        default_factory=list, examples=[[24.0, 20.0]]
+    )
     rope_scale: Optional[float] = Field(
         description="Automatically pulled from the model's config if not present",
         default=None,
@@ -86,16 +88,16 @@ class ModelLoadRequest(BaseModel):
         default=None,
         examples=[1.0],
     )
-    no_flash_attention: bool = False
+    no_flash_attention: Optional[bool] = False
     # low_mem: Optional[bool] = False
-    cache_mode: str = "FP16"
-    chunk_size: int = 2048
+    cache_mode: Optional[str] = "FP16"
+    chunk_size: Optional[int] = 2048
     prompt_template: Optional[str] = None
     num_experts_per_token: Optional[int] = None
     use_cfg: Optional[bool] = None
-    fasttensors: bool = False
+    fasttensors: Optional[bool] = False
     draft: Optional[DraftModelLoadRequest] = None
-    skip_queue: bool = False
+    skip_queue: Optional[bool] = False
 
 
 class ModelLoadResponse(BaseModel):