Tropical badvars natveg nppool 2 pimprove - #67
Open
daliwang wants to merge 5 commits into
Open
Conversation
…tions - Updated `commands.txt` to include new commands for applying bias/scale corrections specifically for Africa and Amazon regions. - Added new configuration files: `finetune_phase2_africa_5p_only.json`, `training_config_africa_5p_box.json`, `training_config_africa_five_p_only.json`, and `training_config_amazon_5p_box.json` to support targeted training and fine-tuning of models focused on phosphorus variables in Africa and Amazon. - Introduced documentation detailing the Africa 5P bias issue and the proposed shrinkage plan in `AFRICA_5P_SHRINKAGE_PLAN.md`. - Enhanced existing documentation to clarify the workflow for training and applying the new configurations, ensuring users can effectively implement the changes. - Added scripts for merging bias-corrected predictions from both regions and updating restart files accordingly.
- Introduced `REPORT_PHASE3_5P_BIAS_CORRECTION_REVIEW.md` detailing the bias correction process for soil phosphorus in the Amazon and Africa regions. - Summarized findings from empirical comparisons between older and newer Phase3 outputs, highlighting improvements and trade-offs in model performance. - Documented related workflows and scripts to enhance user understanding of the bias correction methodology and its implementation. - Provided recommendations for optimizing the bias correction process to achieve more uniform quality across regions.
Add batched inference to avoid GPU OOM on ~260K gridcells, restore export-only CSV generation from saved predictions, and switch restart spatial mapping to KD-tree nearest-neighbor lookup to avoid huge distance matrices.
- Updated `CNP_RATIO_ENFORCEMENT_USAGE.md` to clarify default behavior for N/P derivation during inference and added instructions for skipping derivation in specific scenarios. - Introduced `RERUN_PHASE2_PHASE3_NO_CNP_FIX.md` to document the workflow for rerunning Phase 2 and Phase 3 without CNP ratio fixes, including prerequisites and detailed steps. - Modified `run_inference_all.py` to support optional N/P derivation flags with clearer help descriptions. - Enhanced `apply_5p_africa_shrinkage.py` and `apply_5p_bias_scale_correction.py` scripts to allow for more flexible handling of prediction directories and variable suffixes.
- Updated `ai_predictions_to_netcdf.py` to support an optional `soil_2d_overlay_suffix` for loading overlay CSVs, improving flexibility in handling bias-corrected predictions. - Modified `generate_site_5p_restart_comparison.py` to allow overriding the Phase 3 restart file path via a new command-line argument, enhancing usability for site-specific comparisons.
daliwang
commented
Jun 23, 2026
daliwang
left a comment
daliwang
left a comment
Owner
Author
There was a problem hiding this comment.
merge the new development branch (for trendy case) into main
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
finalized the three phase experiment and validation with the trendy case