codescout-alg-variants

Leon-Leee/codescout-alg-variants

codescout-alg-variants is a machine learning model from Leon-Leee. Use it for the machine learning task on the model card, and read the license before you ship it in a product.

This repository contains checkpoint artifacts for several training variants, selected to enable a fair, quick comparison under similar training compute.

Downloads · 30 days

Access

Public

Updated Mar 4, 2026

Repo size

142 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors80.4 GB · 100%

At a glance

Access: Public
Created: Mar 1, 2026
Updated: Mar 4, 2026
SHA: 9ecbe6ea

Model card

From the Hugging Face model README

Algorithm Config Variants (Step-200 Checkpoints)

This repository contains checkpoint artifacts for several training variants, selected to enable a fair, quick comparison under similar training compute.

W&B:
https://wandb.ai/leon_at_work/soni_ablation_4b/workspace

Included Checkpoints

Variant Folder	Short Description	Checkpoint Step
`Baseline 4B`	Baseline configuration used for comparison	200
`a2_sapo`	Vanilla SAPO (norm_by_std=true + sapo + eps=0.2 + sequence_mean)	200
`a3_dr_grpo_pure`	Dr.GRPO pure (grpo + regular + seq_mean_token_sum_norm + no-std + eps=0.2)	200
`a4_length_norm_sqrt`	GRPO with sqrt length normalization (plus normalized reduction)	200
`a5_sapo_on_sqrt`	SAPO on top of sqrt length-normalized setting	200
`a6_eps_clip_02`	Sqrt length-normalized setting with PPO clip set to 0.2/0.2	200
`a8_dr_grpo_gspo`	Dr.GRPO-style setting with GSPO	200
`a9_dr_grpo_sapo`	Dr.GRPO-style setting with SAPO	200
`a10_grpo_norm_by_std`	GRPO with std-based normalization enabled	200
`a11_vanilla_grpo`	Vanilla GRPO-style setting (`norm_by_std + regular loss + eps=0.2`)	200

Training-Parameter Deltas

Baseline reference (Baseline 4B):
advantage_estimator=grpo, policy_loss_type=gspo, loss_reduction=sequence_mean, eps_clip_low/high=3e-4/4e-4, grpo_norm_by_std=false, reward config includes multilevel_localization_f1_reward + multiturn_reward(minimal_turns=4, maximal_turns=4).

Variant Folder	Parameter Delta vs `Baseline 4B`
`a2_sapo`	`advantage_estimator=grpo`; `policy_loss_type=sapo`; `loss_reduction=sequence_mean`; `eps_clip_low/high=0.2/0.2`; `grpo_norm_by_std=true`
`a3_dr_grpo_pure`	advantage_estimator=grpo; policy_loss_type=regular; loss_reduction=seq_mean_token_sum_norm; eps_clip_low/high=0.2/0.2; grpo_norm_by_std=false
`a4_length_norm_sqrt`	`advantage_estimator=grpo_length_norm_sqrt`; `loss_reduction=seq_mean_token_sum_norm`
`a5_sapo_on_sqrt`	same as `a4_length_norm_sqrt`, plus `policy_loss_type=sapo`
`a6_eps_clip_02`	`a4`-style setup, with `eps_clip_low=0.2`, `eps_clip_high=0.2`
`a8_dr_grpo_gspo`	`advantage_estimator=grpo`; `policy_loss_type=gspo`; `loss_reduction=seq_mean_token_sum_norm`
`a9_dr_grpo_sapo`	`advantage_estimator=grpo`; `policy_loss_type=sapo`; `loss_reduction=seq_mean_token_sum_norm`
`a10_grpo_norm_by_std`	`grpo_norm_by_std=true` (other key settings close to baseline)
`a11_vanilla_grpo`	`grpo_norm_by_std=true`; `policy_loss_type=regular`; `eps_clip_low/high=0.2/0.2`; `loss_reduction=sequence_mean`

Notes on Early Observations

In quick-pass validation (subset-based), these algorithmic variants did not clearly outperform the current baseline yet.
Vanilla GRPO can show higher training reward while not necessarily improving validation performance.
Data prefiltering appears promising from the learning-curve shape, and related runs are still being monitored.

Extra Files

configs/rewards/baseline_4b.yaml is included for reference to the baseline reward configuration.

Created: Mar 1, 2026
Updated: Mar 4, 2026