[0z5a]

Incremental review: Files changed.

Base: 0z5a-branch at 791f101e20db377b28e15f12a00863baaea2605f.

Reject malformed sampled-token log probabilities before appending generated data or calculating rewards. Previously, two token IDs with one probability could be padded with zero and accepted as training data.

Validate non-negative integer token IDs, aligned entries and finite numeric values. Preserve real zeros, finite vLLM sentinels, empty terminal responses, non-trainable tool tokens and multimodal canonical prefixes. Error messages include the local request ID without prompt contents.

The diff contains the parser and two updated expectations in existing tests. Newly added cases were removed.

Validation of the implementation	Result
Publication-checkout rollout checks	99 passed
Repository hooks	All 9 passed
Earlier local/Linux rollout and Sample checks	111 passed
Qwen3-0.6B integration	2 complete RL rounds; 16 samples verified the parser hash
Checkpoint-to-serving reload	Both checkpoints matched all 226 served tensors byte-for-byte

The integration used real generation, response-dependent reward, full backpropagation, AdamW, checkpoint persistence and weight reload. Its local-attention/runtime compatibility changes were outside this patch. Archived checks include the now-removed test harnesses; cleanup preserves production code. This is a correctness fix with no speedup claim.

[read-the-docs-community]

Documentation build overview

📚 vime | 🛠️ Build #34876581 | 📁 Comparing ba45fe8 against latest (ce92eff)

  🔍 Preview build  

45 files changed · ± 45 modified

± Modified

• genindex.html
• index.html
• search.html
• advanced/arch-support-beyond-megatron.html
• advanced/delta-weight-sync.html
• advanced/external-rollout-engines.html
• advanced/fault-tolerance.html
• advanced/low-precision.html
• advanced/megatron-config.html
• advanced/observability.html
• and 35 more...

[gemini-code-assist]

Code Review

This pull request introduces strict validation for vLLM generation metadata (such as token IDs and log probabilities) to prevent silent fallbacks, ensures prompt preparation changes remain local until validation succeeds, and adds unique request IDs to headers and error messages. It also significantly expands the test suite. The reviewer suggests simplifying the validation logic in _inference_generate_tokens_and_logprobs by checking for empty token_ids early to avoid redundant checks and improve clarity.

[gemini-code-assist]

The validation logic for empty token_ids can be significantly simplified. By checking if not token_ids: early, we can immediately return [], [] and avoid redundant checks for logprobs and content being None when token_ids is empty. This also prevents potentially misleading error messages (e.g., raising 'logprobs must be an object for nonempty trainable token_ids' when token_ids is actually empty but logprobs is of an invalid type).

    if not isinstance(choice, dict):
        raise ValueError("choice must be an object")
    token_ids = choice.get("token_ids")
    if not isinstance(token_ids, list) or any(type(token_id) is not int or token_id < 0 for token_id in token_ids):
        raise ValueError("token_ids must be a list of non-negative integers")

    if not token_ids:
        return [], []

    logprobs = choice.get("logprobs")
    if not isinstance(logprobs, dict):
        raise ValueError("logprobs must be an object for nonempty trainable token_ids")
    content = logprobs.get("content")
    if not isinstance(content, list):
        raise ValueError("logprobs.content must be a list")
    if len(content) != len(token_ids):
        raise ValueError(f"token/logprob length mismatch: {len(token_ids)} tokens, {len(content)} entries")

    log_probs = []
    for index, entry in enumerate(content):
        if not isinstance(entry, dict) or "logprob" not in entry:
            raise ValueError(f"missing logprob at token index {index}")
        value = entry["logprob"]
        if isinstance(value, bool) or not isinstance(value, (int, float)):
            raise ValueError(f"non-numeric logprob at token index {index}")
        try:
            value = float(value)
        except OverflowError as exc:
            raise ValueError(f"non-finite logprob at token index {index}") from exc
        if not math.isfinite(value):
            raise ValueError(f"non-finite logprob at token index {index}")
        log_probs.append(value)
    return token_ids, log_probs

[0z5a]

Superseded by #460, now merged into 0z5a-branch. The consolidated implementation and validation are retained there; closing this duplicate PR.