feat: update README.md

code.py

@@ -1,4 +1,3 @@
-
 try:
     from sympy import *
 

configs/config.yaml

@@ -10,19 +10,6 @@ TOTAL_TOKENS: 2048 # if PRIVATE else 512
 
 DEEP: True
 
-code: """Below is a math problem you are to solve (positive numerical answer):
-\"{}\"
-To accomplish this, first determine a sympy-based approach for solving the problem by listing each step to take and what functions need to be called in each step. Be clear so even an idiot can follow your instructions, and remember, your final answer should be positive integer, not an algebraic expression!
-Write the entire script covering all the steps (use comments and document it well) and print the result. After solving the problem, output the final numerical answer within \\boxed{}.
-
-Approach:"""
-
-
-cot: """Below is a math problem you are to solve (positive numerical answer!):
-\"{}\"Analyze this problem and think step by step to come to a solution with programs. After solving the problem, output the final numerical answer within \\boxed{}.\n\n"""
-
-tool_instruction: '\n\nPlease integrate natural language reasoning with programs to solve the above problem, and put your final numerical answer within \\boxed{}.\nNote that the intermediary calculations may be real numbers, but the final numercal answer would always be an integer.'
-
 
 temperature: 0.9
 top_p: 1.0

log.txt

@@ -1 +1,3 @@
 Loading the config file...
+Transformers Version: 4.39.3
+torch.bfloat16

main.py

@@ -10,6 +10,7 @@
 import re
 import math
 import random
+
 from collections import defaultdict
 from collections import Counter
 
@@ -46,50 +47,49 @@ def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor):
         return False
 
 if __name__ == "__main__":
+    import aimo
+    env = aimo.make_env()
+    iter_test = env.iter_test()
+    # # Parse arguments
+    # arg_parser = argparse.ArgumentParser()
+    # arg_parser.add_argument("-c", "--config", default="./configs/config.yaml",
+    #                         help="the config file to be used to run the experiment")
+    # arg_parser.add_argument("--verbose", action='store_true', help="Log also to stdout")
+
+    # args = arg_parser.parse_args()
 
-
-    LOGGER=True
 
-    if LOGGER:
-        log_fn = 'log.txt'
-        if os.path.exists(log_fn):
-            os.remove(log_fn)
-        logger = logging.getLogger()
-        logger.setLevel(logging.INFO)
-        # Create a file handler
-        file_handler = logging.FileHandler(log_fn)
-        file_handler.setLevel(logging.INFO)
-        # Create a stream handler
-        stream_handler = logging.StreamHandler()
-        stream_handler.setLevel(logging.INFO)
-
-        # Add the handlers to the logger
-        logger.addHandler(file_handler)
-        logger.addHandler(stream_handler)
-
-    # Parse arguments
-    arg_parser = argparse.ArgumentParser()
-    arg_parser.add_argument("-c", "--config", default="./configs/config.yaml",
-                            help="the config file to be used to run the experiment")
-    arg_parser.add_argument("--verbose", action='store_true', help="Log also to stdout")
-
-    args = arg_parser.parse_args()
 
+    LOGGER=True
+    log_fn = 'log.txt'
+    if os.path.exists(log_fn):
+        os.remove(log_fn)
+    logger = logging.getLogger()
+    logger.setLevel(logging.INFO)
+    # Create a file handler
+    file_handler = logging.FileHandler(log_fn)
+    file_handler.setLevel(logging.INFO)
+    # Create a stream handler
+    stream_handler = logging.StreamHandler()
+    stream_handler.setLevel(logging.INFO)
+
+    # Add the handlers to the logger
+    logger.addHandler(file_handler)
+    logger.addHandler(stream_handler)
+
+
+    config_fn = './configs/config.yaml'
     # check if the config files exists
-    if not os.path.exists(args.config):
-        logging.info("Config file does not exist: {}".format(args.config))
+    if not os.path.exists(config_fn):
+        logging.info("Config file does not exist: {}".format(config_fn))
         raise SystemExit
 
     # Munchify the dict to access entries with both dot notation and ['name']
     logging.info(f'Loading the config file...')
-    config = yaml.load(open(args.config, "r", encoding='utf-8'), yaml.FullLoader, )
+    config = yaml.load(open(config_fn, "r", encoding='utf-8'), yaml.FullLoader)
     config = munchify(config)
 
 
-    import aimo
-    env = aimo.make_env()
-    iter_test = env.iter_test()
-
 
     import transformers
     logging.info(f"Transformers Version: {transformers.__version__}")
@@ -113,18 +113,14 @@ def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor):
 
     torch.cuda.empty_cache()
     gc.collect()
-
-
-
-
 
     n_repetitions = config.n_repetitions # Original notebook had 22 but times out :(
     TOTAL_TOKENS = config.TOTAL_TOKENS # if PRIVATE else 512
 
     if PRIVATE:
         TIME_LIMIT = 31500
     else:
-        TIME_LIMIT = 31500 # ORIGIN 1
+        TIME_LIMIT = 1 # ORIGIN 1
 
 
 
@@ -217,13 +213,19 @@ def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor):
         logging.info(model.dtype, model.hf_device_map)
 
 
-    code = config.code
+    code =  """Below is a math problem you are to solve (positive numerical answer):
+            \"{}\"
+            To accomplish this, first determine a sympy-based approach for solving the problem by listing each step to take and what functions need to be called in each step. Be clear so even an idiot can follow your instructions, and remember, your final answer should be positive integer, not an algebraic expression!
+            Write the entire script covering all the steps (use comments and document it well) and print the result. After solving the problem, output the final numerical answer within \\boxed{}.
 
+            Approach:"""
 
-    cot = config.cot
+
+    cot = """Below is a math problem you are to solve (positive numerical answer!):
+        \"{}\"Analyze this problem and think step by step to come to a solution with programs. After solving the problem, output the final numerical answer within \\boxed{}.\n\n"""
     promplt_options = [code,cot]
 
-    tool_instruction = config.tool_instruction
+    tool_instruction = '\n\nPlease integrate natural language reasoning with programs to solve the above problem, and put your final numerical answer within \\boxed{}.\nNote that the intermediary calculations may be real numbers, but the final numercal answer would always be an integer.'
 
 
     temperature = config.temperature
@@ -240,6 +242,7 @@ def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor):
     question_type_counts = {}
     starting_counts = (2,3)
 
+
     # LEWIS: I had to invert the loop order because the new API forbids repeated calls on the same problem
     for i, (test, sample_submission) in tqdm(enumerate(iter_test)):
         logging.info(f"Solving problem {i} ...")

test.csv

@@ -1,4 +1,4 @@
-"row_id", ""id","problem","answer"
+"row_id", "id","problem","answer"
 1, "229ee8","Let $k, l > 0$ be parameters. The parabola $y = kx^2 - 2kx + l$ intersects the line $y = 4$ at two points $A$ and $B$. These points are distance 6 apart. What is the sum of the squares of the distances from $A$ and $B$ to the origin?",52
 2, "246d26","Each of the three-digits numbers $111$ to $999$ is coloured blue or yellow in such a way that the sum of any two (not necessarily different) yellow numbers is equal to a blue number. What is the maximum possible number of yellow numbers there can be?",250
 3, "2fc4ad","Let the `sparkle' operation on positive integer $n$ consist of calculating the sum of the digits of $n$ and taking its factorial, e.g. the sparkle of 13 is $4! = 24$. A robot starts with a positive integer on a blackboard, then after each second for the rest of eternity, replaces the number on the board with its sparkle. For some `special' numbers, if they're the first number, then eventually every number that appears will be less than 6. How many such special numbers are there with at most 36 digits?",702