Create get_token_ids.py

get_token_ids.py

@@ -0,0 +1,92 @@
+import tiktoken
+
+# Mapping of model names to their respective encodings
+ENCODINGS = {
+    "gpt-4": tiktoken.get_encoding("cl100k_base"),
+    "gpt-3.5-turbo": tiktoken.get_encoding("cl100k_base"),
+    "gpt-3.5-turbo-0301": tiktoken.get_encoding("cl100k_base"),
+    "text-davinci-003": tiktoken.get_encoding("p50k_base"),
+    "text-davinci-002": tiktoken.get_encoding("p50k_base"),
+    "text-davinci-001": tiktoken.get_encoding("r50k_base"),
+    "text-curie-001": tiktoken.get_encoding("r50k_base"),
+    "text-babbage-001": tiktoken.get_encoding("r50k_base"),
+    "text-ada-001": tiktoken.get_encoding("r50k_base"),
+    "davinci": tiktoken.get_encoding("r50k_base"),
+    "curie": tiktoken.get_encoding("r50k_base"),
+    "babbage": tiktoken.get_encoding("r50k_base"),
+    "ada": tiktoken.get_encoding("r50k_base"),
+}
+
+# Mapping of model names to their respective maximum context lengths
+MAX_LENGTH = {
+    "gpt-4": 8192,
+    "gpt-3.5-turbo": 4096,
+    "gpt-3.5-turbo-0301": 4096,
+    "text-davinci-003": 4096,
+    "text-davinci-002": 4096,
+    "text-davinci-001": 2049,
+    "text-curie-001": 2049,
+    "text-babbage-001": 2049,
+    "text-ada-001": 2049,
+    "davinci": 2049,
+    "curie": 2049,
+    "babbage": 2049,
+    "ada": 2049
+}
+
+def count_tokens(model_name, text):
+    """
+    Count the number of tokens for a given model and text.
+
+    Parameters:
+    - model_name (str): The name of the model.
+    - text (str): The input text.
+
+    Returns:
+    - int: The number of tokens.
+    """
+    if model_name not in ENCODINGS:
+        raise ValueError(f"Model name '{model_name}' not found in encodings.")
+    return len(ENCODINGS[model_name].encode(text))
+
+def get_max_context_length(model_name):
+    """
+    Get the maximum context length for a given model.
+
+    Parameters:
+    - model_name (str): The name of the model.
+
+    Returns:
+    - int: The maximum context length.
+    """
+    if model_name not in MAX_LENGTH:
+        raise ValueError(f"Model name '{model_name}' not found in max length dictionary.")
+    return MAX_LENGTH[model_name]
+
+def get_token_ids_for_text(model_name, text):
+    """
+    Get unique token IDs for a given text using the specified model's encoding.
+
+    Parameters:
+    - model_name (str): The name of the model.
+    - text (str): The input text.
+
+    Returns:
+    - list: A list of unique token IDs.
+    """
+    if model_name not in ENCODINGS:
+        raise ValueError(f"Model name '{model_name}' not found in encodings.")
+    encoded_tokens = ENCODINGS[model_name].encode(text)
+    return list(set(encoded_tokens))
+
+def get_token_ids_for_task_parsing(model_name):
+    """
+    Get unique token IDs for task parsing.
+
+    Parameters:
+    - model_name (str): The name of the model.
+
+    Returns:
+    - list: A list of unique token IDs for task parsing.
+    """
+    text = '''{"task": "text-classification", "token-classification", "text2text-generation", "summarization", "translation", "question-answering", "conversational", "text-generation", "sentence-similarity", "tabular-classification", "object-detection", "image-classification", "image-to-image", "image-to-text", "text-to-image", "visual-question-answering", "document-question-answering", "image-segmentation", "text-to-speech", "text-to-video", "automatic-speech-recognition", "audio-to-audio", "audio-classification", "canny-control", "hed-control", "mlsd-control", "normal-control", "openpose-control", "canny-text-to-image", "depth-text-to-image", "hed-text-to-image", "mlsd-text-to-image", "normal-text-to-image", "openpose-text