diogoantunes commited on
Commit
894e783
·
verified ·
1 Parent(s): 98d6922

Document examples and validation; align MIT copyright contacts

Browse files
Files changed (5) hide show
  1. LICENSE +3 -3
  2. README.md +15 -2
  3. SHA256SUMS +4 -4
  4. SOURCE_LICENSE +3 -3
  5. artifact_manifest.json +12 -5
LICENSE CHANGED
@@ -1,8 +1,8 @@
1
  MIT License
2
 
3
- Copyright (c) 2026 Diogo Antunes
4
- Copyright (c) 2026 ValueAI
5
- Copyright (c) 2026 Keplen
6
 
7
  Permission is hereby granted, free of charge, to any person obtaining a copy
8
  of this software and associated documentation files (the "Software"), to deal
 
1
  MIT License
2
 
3
+ Copyright (c) 2026 Diogo Antunes <contact@diogoantunes.ai>
4
+ Copyright (c) 2026 ValueAI <contact@valuealigned.ai>
5
+ Copyright (c) 2026 Keplen <contact@keplen.ai>
6
 
7
  Permission is hereby granted, free of charge, to any person obtaining a copy
8
  of this software and associated documentation files (the "Software"), to deal
README.md CHANGED
@@ -10,7 +10,6 @@ tags:
10
  - embodied-control
11
  - research
12
  - study-artifacts
13
- - unpublished-manuscript
14
  - npz
15
  ---
16
 
@@ -88,6 +87,20 @@ For the BF16 export, download `weights_bf16.npz` instead and use `receipt["weigh
88
 
89
  Use `models/x1b_world.py` for the core, `models/x1b_jepa.py` for latent prediction and readouts, and the matching planner/evaluation modules in the pinned source. `scripts/eval/x1b_infer.py` can load local NPZ checkpoints through its `load_verified_checkpoint` function; supply this artifact directory so the Wave 5 receipt and configuration are used.
90
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
91
  ## Recorded results for this checkpoint
92
 
93
  These values come directly from this run's `receipt.json`. They are **single-seed results**, not the manuscript's cohort mean ± standard deviation. The upload does not claim that the quantized export has received an additional independent evaluation.
@@ -115,7 +128,7 @@ The checkpoint is intended for research and reproducibility. The upload contains
115
  - [SLWM source repository](https://github.com/ValueAlignedAI/slwm/tree/cbb2b38794359adfefe07b19e9b7caa113b25b47)
116
  - [Original experiment records](https://github.com/ValueAlignedAI/slwm/tree/cbb2b38794359adfefe07b19e9b7caa113b25b47/experiments/multimodal/x1b/spatiotemporal-jepa-1b-20261111/runs_wave5/x1b-w5-slwm-1b-s2001)
117
 
118
- These study artifacts, including the BF16 and INT8 weights, use the **MIT license**, as specified by the author. The complete `LICENSE` and `SOURCE_LICENSE` files match the [corrected GitHub project license](https://github.com/ValueAlignedAI/slwm/blob/9c0aa3fb333fa9b099e2ceb4af976970213788ef/LICENSE). The copyright owners are **Diogo Antunes, ValueAI, and Keplen**.
119
 
120
  ## Citation
121
 
 
10
  - embodied-control
11
  - research
12
  - study-artifacts
 
13
  - npz
14
  ---
15
 
 
87
 
88
  Use `models/x1b_world.py` for the core, `models/x1b_jepa.py` for latent prediction and readouts, and the matching planner/evaluation modules in the pinned source. `scripts/eval/x1b_infer.py` can load local NPZ checkpoints through its `load_verified_checkpoint` function; supply this artifact directory so the Wave 5 receipt and configuration are used.
89
 
90
+ ## Run an example and validation checks
91
+
92
+ Use the [runnable examples and validation-set guide](https://github.com/ValueAlignedAI/slwm/blob/main/docs/research/x1b_study_artifacts_and_validation.md) from the SLWM source repository. It explains the development and held-out data splits, how to verify local data-bank checksums, and which recorded inputs are not included with these checkpoints.
93
+
94
+ From a current checkout of [the source repository](https://github.com/ValueAlignedAI/slwm), after installing the guide's Python dependencies:
95
+
96
+ ```sh
97
+ python scripts/eval/x1b_study_examples.py \
98
+ --mode example --repo-id valueai-org/SLWM-X1B --format int8 \
99
+ --output /tmp/slwm-x1b-example.json
100
+ ```
101
+
102
+ This is a generated, deterministic functional example using the selected trained checkpoint; it is not a benchmark score. The guide provides BF16 commands for the original evaluation functions and the exact bank paths and split names for validation.
103
+
104
  ## Recorded results for this checkpoint
105
 
106
  These values come directly from this run's `receipt.json`. They are **single-seed results**, not the manuscript's cohort mean ± standard deviation. The upload does not claim that the quantized export has received an additional independent evaluation.
 
128
  - [SLWM source repository](https://github.com/ValueAlignedAI/slwm/tree/cbb2b38794359adfefe07b19e9b7caa113b25b47)
129
  - [Original experiment records](https://github.com/ValueAlignedAI/slwm/tree/cbb2b38794359adfefe07b19e9b7caa113b25b47/experiments/multimodal/x1b/spatiotemporal-jepa-1b-20261111/runs_wave5/x1b-w5-slwm-1b-s2001)
130
 
131
+ These study artifacts, including the BF16 and INT8 weights, use the **MIT license**, as specified by the author. The complete `LICENSE` and `SOURCE_LICENSE` files match the [corrected GitHub project license](https://github.com/ValueAlignedAI/slwm/blob/d1abfc031d9d0c54f511e9a4eb324dfc0b57a08d/LICENSE). The copyright owners are **Diogo Antunes** ([contact@diogoantunes.ai](mailto:contact@diogoantunes.ai)), **ValueAI** ([contact@valuealigned.ai](mailto:contact@valuealigned.ai)), and **Keplen** ([contact@keplen.ai](mailto:contact@keplen.ai)).
132
 
133
  ## Citation
134
 
SHA256SUMS CHANGED
@@ -1,7 +1,7 @@
1
- f65b241f97baf094a332af0c80f7fa96ddfe4b2b83f9d876b27d3ed31d5e9615 LICENSE
2
- 1032c7c497b7fc6e4cd95502f78c6fcf6a7ae8630311c573b64d44844accbc09 README.md
3
- f65b241f97baf094a332af0c80f7fa96ddfe4b2b83f9d876b27d3ed31d5e9615 SOURCE_LICENSE
4
- 81a9a08813c13542ce37051d8003def89eb633445a7c8276fdff8a130e09b995 artifact_manifest.json
5
  2e31e53c6431657da058c6293d3c7b948dfb5b221fe5c4fde1c08ea8c5bd2c00 config.json
6
  ef8878a15da85fd1e14724e6ad50b6d0a6e6c05f1e1aa673c910f44ff576771b eval_metrics.json
7
  27686dfc84db3c70b74999e2ee665df6833b99e380df93d3573c857c9c240ca8 load_checkpoint.py
 
1
+ fc6321c8fbaeec919739657e6a4ac859bc1452f6b480b4dc1e3e3de74afd6010 LICENSE
2
+ ca2264defc08ffee77d59abb65db4f3d48087b3892c641886508340427d804fc README.md
3
+ fc6321c8fbaeec919739657e6a4ac859bc1452f6b480b4dc1e3e3de74afd6010 SOURCE_LICENSE
4
+ 74bc9e5c39b5f0eece6d20ec908693cb9a5942a3dfce5149f625a661cd22267b artifact_manifest.json
5
  2e31e53c6431657da058c6293d3c7b948dfb5b221fe5c4fde1c08ea8c5bd2c00 config.json
6
  ef8878a15da85fd1e14724e6ad50b6d0a6e6c05f1e1aa673c910f44ff576771b eval_metrics.json
7
  27686dfc84db3c70b74999e2ee665df6833b99e380df93d3573c857c9c240ca8 load_checkpoint.py
SOURCE_LICENSE CHANGED
@@ -1,8 +1,8 @@
1
  MIT License
2
 
3
- Copyright (c) 2026 Diogo Antunes
4
- Copyright (c) 2026 ValueAI
5
- Copyright (c) 2026 Keplen
6
 
7
  Permission is hereby granted, free of charge, to any person obtaining a copy
8
  of this software and associated documentation files (the "Software"), to deal
 
1
  MIT License
2
 
3
+ Copyright (c) 2026 Diogo Antunes <contact@diogoantunes.ai>
4
+ Copyright (c) 2026 ValueAI <contact@valuealigned.ai>
5
+ Copyright (c) 2026 Keplen <contact@keplen.ai>
6
 
7
  Permission is hereby granted, free of charge, to any person obtaining a copy
8
  of this software and associated documentation files (the "Software"), to deal
artifact_manifest.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "repo_id": "valueai-org/SLWM-X1B",
3
- "title": "SLWM-X1B \u2014 Study artifacts (Wave 5)",
4
  "wave": 5,
5
  "run_tag": "x1b-w5-slwm-1b-s2001",
6
  "seed": 2001,
@@ -65,12 +65,19 @@
65
  "formal_model_release": false,
66
  "purpose": "Reference artifacts accompanying the unpublished manuscript draft v0.3.",
67
  "license": "mit",
68
- "license_source": "https://github.com/ValueAlignedAI/slwm/blob/9c0aa3fb333fa9b099e2ceb4af976970213788ef/LICENSE",
69
- "license_sha256": "f65b241f97baf094a332af0c80f7fa96ddfe4b2b83f9d876b27d3ed31d5e9615",
70
- "license_source_commit": "9c0aa3fb333fa9b099e2ceb4af976970213788ef",
71
  "copyright_owners": [
72
  "Diogo Antunes",
73
  "ValueAI",
74
  "Keplen"
75
- ]
 
 
 
 
 
 
 
76
  }
 
1
  {
2
  "repo_id": "valueai-org/SLWM-X1B",
3
+ "title": "SLWM-X1B — Study artifacts (Wave 5)",
4
  "wave": 5,
5
  "run_tag": "x1b-w5-slwm-1b-s2001",
6
  "seed": 2001,
 
65
  "formal_model_release": false,
66
  "purpose": "Reference artifacts accompanying the unpublished manuscript draft v0.3.",
67
  "license": "mit",
68
+ "license_source": "https://github.com/ValueAlignedAI/slwm/blob/d1abfc031d9d0c54f511e9a4eb324dfc0b57a08d/LICENSE",
69
+ "license_sha256": "fc6321c8fbaeec919739657e6a4ac859bc1452f6b480b4dc1e3e3de74afd6010",
70
+ "license_source_commit": "d1abfc031d9d0c54f511e9a4eb324dfc0b57a08d",
71
  "copyright_owners": [
72
  "Diogo Antunes",
73
  "ValueAI",
74
  "Keplen"
75
+ ],
76
+ "copyright_contacts": {
77
+ "Diogo Antunes": "contact@diogoantunes.ai",
78
+ "ValueAI": "contact@valuealigned.ai",
79
+ "Keplen": "contact@keplen.ai"
80
+ },
81
+ "validation_guide": "https://github.com/ValueAlignedAI/slwm/blob/main/docs/research/x1b_study_artifacts_and_validation.md",
82
+ "organization_card": "https://huggingface.co/valueai-org"
83
  }