Training a small model to write better OCaml with RLVR and GRPO
Kiran Gangadharan explores the training of a small language model to improve OCaml code generation using RLVR and GRPO techniques. The experiment involved training a 1.5B model on a dataset derived from public GitHub repositories, focusing on OCaml's unique syntax. Key aspects included defining constraints for local inference and using a graduated reward system to enhance the model's learning process.
- ▪The model was trained on a small dataset of programming problems adapted to OCaml.
- ▪A single rented GPU was used for training, with LoRA to reduce memory requirements.
- ▪The training loop utilized Hugging Face's trl library for GRPO integration.
Hacker News (Newest) files mainly under programming. We currently carry 5,306 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | nilenso blog |
| Canonical URL | https://blog.nilenso.com/blog/2026/05/18/training-a-small-model-to-write-better-ocaml-with-rlvr-and-grpo/ |
| Publication time | Wed, 20 May 2026 18:27:28 +0000 |
| Retrieval time | 2026-05-20T18:35:02.972Z |
| Last seen | 2026-05-20T18:35:02.972Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | CMidMAax5BqL |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Kiran Gangadharan Training a small model to write better OCaml with RLVR and GRPO 18 May 2026 For a while now, I’ve been interested in exploring the capabilities of small language models. When my colleague Atharva introduced me to RLVR and GRPO for doing RL training without a human feedback loop, I wanted to know more. In the previous post, we explored the workings of RLVR and GRPO. In this post, I’ll walk through a code-generation experiment where I trained a small 1.5B model with GRPO, improved its ability to generate correct and valid OCaml code, and share what I learned along the way.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at nilenso blog.
Discussion
0 commentsMore from nilenso blog
