Upload code/set1_polypythia.py with huggingface_hub

code/set1_polypythia.py

@@ -9,14 +9,17 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 
 ap = argparse.ArgumentParser()
 ap.add_argument("--size", default="14m")
+ap.add_argument("--base", default=None, help="repo stem for tokenizer/pseudo-base; defaults to pythia-<size>")
 ap.add_argument("--seeds", default="1,2,3,4,5,6,7,8,9")
 ap.add_argument("--blocks", type=int, default=48)
 ap.add_argument("--bs", type=int, default=16)
 ap.add_argument("--barrier_n", type=int, default=7)
 ap.add_argument("--acts_rows", type=int, default=2048)
+ap.add_argument("--tag", default="set1")
 A = ap.parse_args()
+BASE_STEM = A.base or f"pythia-{A.size}"
 SEEDS = [int(s) for s in A.seeds.split(",")]
-OUT = f"/root/compose-audit/results/set1_{A.size}.jsonl"
+OUT = f"/root/compose-audit/results/{A.tag}_{A.size}.jsonl"
 DEV = "cuda"
 
 
@@ -73,7 +76,7 @@ def align_full(sd_a, sd_b, d, acts_a, acts_b, n_heads, nlayer, method):
 
 
 # ------------------------------------------------------------------ setup
-tok = AutoTokenizer.from_pretrained(f"EleutherAI/pythia-{A.size}")
+tok = AutoTokenizer.from_pretrained(f"EleutherAI/{BASE_STEM}")
 lines = flores_lines("eng_Latn")
 blocks = make_blocks(tok, lines, block=512, max_blocks=A.blocks)
 log(f"size={A.size} blocks={tuple(blocks.shape)} tokens={blocks.numel()}")
@@ -101,10 +104,10 @@ for s in SEEDS:
 
 BASE = None
 try:
-    mb = AutoModelForCausalLM.from_pretrained(f"EleutherAI/pythia-{A.size}", dtype=torch.float32).to(DEV).eval()
+    mb = AutoModelForCausalLM.from_pretrained(f"EleutherAI/{BASE_STEM}", dtype=torch.float32).to(DEV).eval()
     BASE = sd_np(mb); BASE_NLL = nll_nats(mb, blocks, DEV, bs=A.bs)
     del mb; torch.cuda.empty_cache()
-    log(f"  base(pythia-{A.size}, NOT a shared ancestor of the seeds) nll={BASE_NLL:.4f}")
+    log(f"  base({BASE_STEM}, NOT a shared ancestor of the seeds) nll={BASE_NLL:.4f}")
 except Exception as e:
     log("base load failed:", e)