update

README.md

@@ -310,22 +310,31 @@ python generation/main.py fit --config generation/confs/cli_lean4_novel_premises
 
 After the tactic generator is trained, we combine it with best-first search to prove theorems by interacting with Lean.
 
-For models without retrieval, run:
+The evaluation script takes Hugging Face model checkpoints (either local or remote) as input. For remote models, you can simply use their names, e.g., [kaiyuy/leandojo-lean4-tacgen-byt5-small](https://huggingface.co/kaiyuy/leandojo-lean4-tacgen-byt5-small). For locally trained models, you first need to convert them from PyTorch Ligthning checkpoints to Hugging Face checkpoints:
 ```bash
-python prover/evaluate.py --data-path data/leandojo_benchmark_4/random/ --ckpt_path $PATH_TO_MODEL_CHECKPOINT --split test --num-workers 5 --num-gpus 1
-python prover/evaluate.py --data-path data/leandojo_benchmark_4/novel_premises/ --ckpt_path $PATH_TO_MODEL_CHECKPOINT --split test --num-workers 5 --num-gpus 1
+python scripts/convert_checkpoint.py generator --src $PATH_TO_GENERATOR_CHECKPOINT --dst ./leandojo-lean4-tacgen-byt5-small
+python scripts/convert_checkpoint.py retriever --src $PATH_TO_RETRIEVER_CHECKPOINT --dst ./leandojo-lean4-retriever-byt5-small
 ```
+, where `PATH_TO_GENERATOR_CHECKPOINT` and `PATH_TO_RETRIEVER_CHECKPOINT` are PyTorch Ligthning checkpoints produced by the training script.
 
-For models with retrieval, first use the retriever to index the corpus (pre-computing the embeddings of all premises):
+
+To evaluate the model without retrieval, run (using the `random` data split as example):
+```bash
+python prover/evaluate.py --data-path data/leandojo_benchmark_4/random/ --gen_ckpt_path ./leandojo-lean4-tacgen-byt5-small --split test --num-workers 5 --num-gpus 1
+```
+You may tweak `--num-workers` and `--num-gpus` to fit your hardware.
+
+
+For the model with retrieval, first use the retriever to index the corpus (pre-computing the embeddings of all premises):
 ```bash
-python retrieval/index.py --ckpt_path $PATH_TO_RETRIEVER_CHECKPOINT --corpus-path data/leandojo_benchmark_4/corpus.jsonl --output-path $PATH_TO_INDEXED_CORPUS
-# Do it separately for two data splits.
+python retrieval/index.py --ckpt_path ./leandojo-lean4-retriever-byt5-small --corpus-path data/leandojo_benchmark_4/corpus.jsonl --output-path $PATH_TO_INDEXED_CORPUS
 ```
+It saves the indexed corpurs as a pickle file to `PATH_TO_INDEXED_CORPUS`.
 
 Then, run:
 ```bash
-python prover/evaluate.py --data-path data/leandojo_benchmark_4/random/  --ckpt_path $PATH_TO_REPROVER_CHECKPOINT --indexed-corpus-path $PATH_TO_INDEXED_CORPUS --split test --num-cpus 8 --with-gpus
-# Do it separately for two data splits.
+python scripts/convert_checkpoint.py generator --src $PATH_TO_REPROVER_CHECKPOINT --dst ./leandojo-lean4-retriever-tacgen-byt5-small
+python prover/evaluate.py --data-path data/leandojo_benchmark_4/random/  --gen_ckpt_path ./leandojo-lean4-retriever-tacgen-byt5-small --ret_ckpt_path ./leandojo-lean4-retriever-byt5-small --indexed-corpus-path $PATH_TO_INDEXED_CORPUS --split test --num-workers 5 --num-gpus 1
 ```
 
 

prover/proof_search.py

@@ -65,6 +65,7 @@ def __init__(
         debug: bool,
     ) -> None:
         self.tac_gen = tac_gen
+        self.tac_gen.initialize()
         self.timeout = timeout
         self.num_sampled_tactics = num_sampled_tactics
         self.debug = debug
@@ -309,7 +310,6 @@ def __init__(
         num_sampled_tactics: int,
         debug: bool,
     ) -> None:
-        tac_gen.initialize()
         self.prover = BestFirstSearchProver(
             tac_gen,
             timeout,

retrieval/index.py

@@ -35,7 +35,7 @@ def main() -> None:
     model.reindex_corpus(batch_size=args.batch_size)
 
     pickle.dump(
-        IndexedCorpus(model.corpus, model.corpus_embeddings.cpu()),
+        IndexedCorpus(model.corpus, model.corpus_embeddings.to(torch.float32).cpu()),
         open(args.output_path, "wb"),
     )
     logger.info(f"Indexed corpus saved to {args.output_path}")

scripts/stats.py

@@ -1,17 +1,36 @@
+import re
 import sys
+import numpy as np
 from glob import glob
+from loguru import logger
+import matplotlib.pyplot as plt
 
+total_time = []
+TOTAL_TIME_REGEX = re.compile(r"total_time=(?P<time>.+?),")
 
 for filename in glob(sys.argv[1]):
-    print(filename)
+    logger.info(filename)
     num_total = num_correct = 0
     for line in open(filename):
         if "SearchResult" in line:
             num_total += 1
             if "Proved" in line:
                 num_correct += 1
+                total_time.append(float(TOTAL_TIME_REGEX.search(line)["time"]))
 
     if num_total == 0:
-        print("N/A")
+        logger.info("Pass@1: N/A")
     else:
-        print(f"{num_correct} / {num_total} = {num_correct / num_total}")
+        logger.info(f"Pass@1: {num_correct} / {num_total} = {num_correct / num_total}")
+
+logger.info(f"Average time: {np.mean(total_time)}")
+
+total_time.sort()
+x = []
+y = []
+for i, t in enumerate(total_time):
+    x.append(t)
+    y.append(i + 1)
+plt.scatter(x, y)
+plt.savefig("stats.pdf")
+logger.info("Figure saved to stats.pdf")