training: restructure as vLLM plugin package

- Convert to installable package with entry points for vLLM auto-discovery
- Add checkpoint_sync.py: Python replacement for Rust checkpoint binary
  - Block-level diffing of safetensors files (4KB blocks)
  - vLLM→HF weight name conversion built-in
  - Scheduled 10min after training jobs (batched)
- API change: /train now takes raw token IDs (context_ids + continuation_ids)
  - No tokenizer on training side, client owns tokenization
- Remove superseded code: standalone scripts, Rust binary, tokenizer helpers

Install: pip install -e ./training
Then vLLM auto-loads via entry point.

Co-Authored-By: Proof of Concept <poc@bcachefs.org>

training/vllm_export_hook.py

@@ -1,76 +0,0 @@
-"""Monkey-patch vLLM to export weight IPC handles on startup.
-
-Usage — add to start_vllm.sh BEFORE the vllm serve command:
-
-    export VLLM_PLUGINS=vllm_export_hook
-    vllm serve Qwen/Qwen3.5-27B ...
-
-Or use Python to launch vLLM with the hook:
-
-    python3 -c "
-    import vllm_export_hook  # installs the patch
-    from vllm.entrypoints.openai.api_server import run_server
-    run_server(...)
-    "
-
-The hook patches vLLM's model runner to export IPC handles after
-model loading completes. The handles are saved to a file that the
-Apollo training process reads.
-"""
-
-import atexit
-import torch
-from pathlib import Path
-
-HANDLE_PATH = "/tmp/vllm_weight_handles.pt"
-
-
-def export_model_weights(model):
-    """Export CUDA IPC handles for all model parameters."""
-    from torch.multiprocessing.reductions import reduce_tensor
-
-    handles = {}
-    total_bytes = 0
-
-    for name, param in model.named_parameters():
-        if param.device.type != 'cuda':
-            continue
-        handle = reduce_tensor(param.data)
-        handles[name] = {
-            'handle': handle,
-            'shape': list(param.shape),
-            'dtype': str(param.dtype),
-        }
-        total_bytes += param.nelement() * param.element_size()
-
-    torch.save(handles, HANDLE_PATH)
-    print(f"[apollo] Exported {len(handles)} weight handles "
-          f"({total_bytes / 1e9:.1f} GB) to {HANDLE_PATH}")
-
-
-def _patch_model_runner():
-    """Patch gpu_worker to export handles after model loading.
-
-    vLLM loads the model in a subprocess (EngineCore_DP0), so we
-    can't patch from the parent. Instead, patch the worker's
-    init_device or load_model at the module level — the subprocess
-    imports the same modules.
-    """
-    from vllm.v1.worker import gpu_worker
-
-    original_load = gpu_worker.Worker.load_model
-
-    def patched_load(self, *args, **kwargs):
-        result = original_load(self, *args, **kwargs)
-        try:
-            export_model_weights(self.model_runner.model)
-        except Exception as e:
-            print(f"[apollo] Failed to export weights: {e}")
-        return result
-
-    gpu_worker.Worker.load_model = patched_load
-    print("[apollo] Weight export hook installed")
-
-
-# Auto-install when imported
-_patch_model_runner()