Files
frankenstallm/source/scripts/migrate_qkv_checkpoint.py
ModelHub XC d4abdb70fa 初始化项目,由ModelHub XC社区提供模型
Model: pathcosmos/frankenstallm
Source: Original Platform
2026-07-14 04:21:16 +08:00

231 lines
8.3 KiB
Python
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

#!/usr/bin/env python3
"""Migrate checkpoint from separate Q/K/V projections to fused QKV.
Usage:
python3 scripts/migrate_qkv_checkpoint.py <checkpoint_dir>
Migrates both model.pt AND optimizer.pt:
- model.pt: q_proj/k_proj/v_proj weights → qkv_proj weight
- optimizer.pt: exp_avg/exp_avg_sq states fused, param indices re-mapped
The concatenation order is [Q ; K ; V] along the output (dim-0) axis,
which matches the split in MultiHeadAttention.forward:
q, k, v = qkv.split([_q_dim, _kv_dim, _kv_dim], dim=-1)
Optimizer layout (group 0 = weight_decay, per layer × 28):
[i*6+0] q_proj.weight [3072, 3072]
[i*6+1] k_proj.weight [1024, 3072]
[i*6+2] v_proj.weight [1024, 3072]
[i*6+3] out_proj.weight [3072, 3072]
[i*6+4] fc1_weight [16384, 3072]
[i*6+5] fc2_weight [3072, 8192]
After fusion: indices 0,1,2 → single qkv_proj → 4 params per layer.
"""
import sys
import torch
from pathlib import Path
N_LAYERS = 28
OLD_PARAMS_PER_LAYER = 6 # q, k, v, out, fc1, fc2
NEW_PARAMS_PER_LAYER = 4 # qkv, out, fc1, fc2
def migrate_model(state: dict) -> dict:
"""Fuse Q/K/V projection weights into QKV in model state dict."""
new_state: dict = {}
layers_done: set = set()
for key, val in state.items():
if ".q_proj." not in key and ".k_proj." not in key and ".v_proj." not in key:
new_state[key] = val
continue
if ".q_proj." not in key:
continue
prefix = key.rsplit(".", 2)[0]
suffix = key.rsplit(".", 1)[-1]
tag = (prefix, suffix)
if tag in layers_done:
continue
layers_done.add(tag)
q_key = f"{prefix}.q_proj.{suffix}"
k_key = f"{prefix}.k_proj.{suffix}"
v_key = f"{prefix}.v_proj.{suffix}"
missing = [k for k in (q_key, k_key, v_key) if k not in state]
if missing:
raise KeyError(f"Expected keys not found in checkpoint: {missing}")
q_w, k_w, v_w = state[q_key], state[k_key], state[v_key]
fused = torch.cat([q_w, k_w, v_w], dim=0)
fused_key = f"{prefix}.qkv_proj.{suffix}"
new_state[fused_key] = fused
print(f" Fused {fused_key}: {list(fused.shape)}"
f" (q={list(q_w.shape)}, k={list(k_w.shape)}, v={list(v_w.shape)})")
leaked = [k for k in new_state if ".q_proj." in k or ".k_proj." in k or ".v_proj." in k]
if leaked:
raise RuntimeError(f"BUG: old projection keys still present: {leaked}")
return new_state
def migrate_optimizer(opt_state: dict) -> dict:
"""Fuse optimizer states for Q/K/V → QKV and re-index parameters.
The optimizer has 2 param groups:
Group 0 (weight_decay): 168 = 28 layers × 6 (q,k,v,out,fc1,fc2)
Group 1 (no weight_decay): 58 = norms + embedding
We fuse q,k,v entries in group 0 (indices i*6+0,1,2 → one entry per layer).
Group 0 shrinks from 168 to 112 (28 layers × 4 params).
Group 1 stays at 58. Total: 170.
"""
old_state = opt_state["state"]
old_groups = opt_state["param_groups"]
group0_count = len(old_groups[0]["params"])
expected_g0 = N_LAYERS * OLD_PARAMS_PER_LAYER
if group0_count != expected_g0:
raise ValueError(
f"Group 0 has {group0_count} params, expected {expected_g0}. "
f"Cannot auto-detect QKV layout."
)
# Validate shapes for first layer
shapes = []
for j in range(OLD_PARAMS_PER_LAYER):
idx = old_groups[0]["params"][j]
shapes.append(list(old_state[idx]["exp_avg"].shape))
expected_shapes = [[3072, 3072], [1024, 3072], [1024, 3072],
[3072, 3072], [16384, 3072], [3072, 8192]]
if shapes != expected_shapes:
raise ValueError(
f"Layer 0 shapes {shapes} don't match expected {expected_shapes}. "
f"Cannot auto-detect QKV layout."
)
print(f" Shape validation passed for layer 0.")
new_state_entries = {}
new_idx = 0
# --- Group 0: fuse q/k/v per layer ---
for layer_i in range(N_LAYERS):
base = layer_i * OLD_PARAMS_PER_LAYER
q_opt_idx = old_groups[0]["params"][base + 0]
k_opt_idx = old_groups[0]["params"][base + 1]
v_opt_idx = old_groups[0]["params"][base + 2]
q_entry = old_state[q_opt_idx]
k_entry = old_state[k_opt_idx]
v_entry = old_state[v_opt_idx]
# Fuse QKV
fused_entry = {"step": q_entry["step"]}
for field in ["exp_avg", "exp_avg_sq"]:
if field in q_entry:
fused_entry[field] = torch.cat(
[q_entry[field], k_entry[field], v_entry[field]], dim=0
)
new_state_entries[new_idx] = fused_entry
if layer_i == 0:
print(f" Layer 0 QKV fused: exp_avg {list(fused_entry['exp_avg'].shape)}")
new_idx += 1
# Copy remaining params (out, fc1, fc2)
for offset in [3, 4, 5]:
opt_idx = old_groups[0]["params"][base + offset]
new_state_entries[new_idx] = old_state[opt_idx]
new_idx += 1
new_group0_count = new_idx # should be N_LAYERS * NEW_PARAMS_PER_LAYER = 112
print(f" Group 0: {group0_count}{new_group0_count} params")
# --- Group 1: copy as-is (norms, embedding — no QKV) ---
group1_count = len(old_groups[1]["params"])
for j in range(group1_count):
opt_idx = old_groups[1]["params"][j]
if opt_idx in old_state:
new_state_entries[new_idx] = old_state[opt_idx]
new_idx += 1
print(f" Group 1: {group1_count}{group1_count} params (unchanged)")
# Build new param_groups
new_groups = []
g0 = {k: v for k, v in old_groups[0].items() if k != "params"}
g0["params"] = list(range(0, new_group0_count))
new_groups.append(g0)
g1 = {k: v for k, v in old_groups[1].items() if k != "params"}
g1["params"] = list(range(new_group0_count, new_group0_count + group1_count))
new_groups.append(g1)
total = new_group0_count + group1_count
print(f" Total: {len(old_state)}{total} optimizer params")
return {"state": new_state_entries, "param_groups": new_groups}
def migrate(ckpt_dir: Path) -> None:
model_path = ckpt_dir / "model.pt"
opt_path = ckpt_dir / "optimizer.pt"
if not model_path.exists():
raise FileNotFoundError(f"model.pt not found in {ckpt_dir}")
# --- Model migration ---
print(f"[1/2] Migrating model weights from {model_path} ...")
state = torch.load(model_path, map_location="cpu", weights_only=True)
has_old = any(".q_proj." in k for k in state)
has_new = any(".qkv_proj." in k for k in state)
if has_new and not has_old:
print(" Model already migrated. Skipping.")
elif has_old:
new_model_state = migrate_model(state)
torch.save(new_model_state, model_path)
print(f" Model saved.")
else:
raise RuntimeError("Model state has neither q_proj nor qkv_proj keys!")
# --- Optimizer migration ---
if opt_path.exists():
print(f"\n[2/2] Migrating optimizer states from {opt_path} ...")
opt = torch.load(opt_path, map_location="cpu", weights_only=True)
# Check if already migrated
total_params = sum(len(pg["params"]) for pg in opt["param_groups"])
expected_old = N_LAYERS * OLD_PARAMS_PER_LAYER + 58 # 168 + 58 = 226
expected_new = N_LAYERS * NEW_PARAMS_PER_LAYER + 58 # 112 + 58 = 170
if total_params == expected_old:
opt_backup = ckpt_dir / "optimizer.pt.backup_pre_qkv"
if not opt_backup.exists():
torch.save(opt, opt_backup)
print(f" Backup: {opt_backup}")
new_opt = migrate_optimizer(opt)
torch.save(new_opt, opt_path)
print(f" Optimizer saved.")
elif total_params == expected_new:
print(f" Optimizer already migrated ({total_params} params). Skipping.")
else:
print(f" [WARN] Unexpected param count {total_params} "
f"(expected old={expected_old} or new={expected_new}). "
f"Deleting optimizer.pt — optimizer will restart fresh.")
opt_path.unlink()
else:
print("\n[2/2] No optimizer.pt found. Optimizer will restart fresh.")
print("\nMigration complete!")
if __name__ == "__main__":
if len(sys.argv) != 2:
print(__doc__)
sys.exit(1)
migrate(Path(sys.argv[1]))