Finalize Pollock 1.3 release metadata

README.md

@@ -123,11 +123,11 @@ Model używa standardowego `GPT2LMHeadModel`; `trust_remote_code=True` nie jest
 
 ### Stałe próbki inferencji
 
-Wspólny zestaw [`fixed-sampling-v1`](./inference-samples/README.md) pokazuje te same cztery prompty wygenerowane przez r001-r005. Opublikowane rewizje są ładowane z pełnych, niezmiennych SHA commitów; przedpublikacyjny artefakt r005 jest ładowany lokalnie. SHA-256 każdego pliku z wagami jest sprawdzane przed inferencją. Referencyjny protokół używa macOS 26.2 na arm64, CPU, float32, jednego wątku, Transformers 5.15.1, seed 1337 resetowanego dla każdego promptu, temperature 0.7, top-k 50 i limitu 100 nowych tokenów.
+Wspólny zestaw [`fixed-sampling-v1`](./inference-samples/README.md) pokazuje te same cztery prompty wygenerowane przez r001-r005. Wszystkie rewizje są ładowane z pełnych, niezmiennych SHA commitów, a SHA-256 każdego pliku z wagami jest sprawdzane przed inferencją. Referencyjny protokół używa macOS 26.2 na arm64, CPU, float32, jednego wątku, Transformers 5.15.1, seed 1337 resetowanego dla każdego promptu, temperature 0.7, top-k 50 i limitu 100 nowych tokenów.
 
 | Rewizja | Wydanie | Wagi | Historia |
 |---|---|---|---|
-| **r005** | **Pollock 1.3 (pakiet do recenzji)** | lokalny artefakt; commit pending | [`r005.md`](./training-history/r005.md) |
+| **r005** | **Pollock 1.3** | [`e780025`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/e780025f15e06bc3765a74d973906eb8a11c022c) | [`r005.md`](./training-history/r005.md) |
 | r004 | Pollock 1.2 | [`30feb81`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/30feb81e5097eef13b939e1827d98e0458bf602d) | [`r004.md`](./training-history/r004.md) |
 | r003 | Pollock 1.1 | [`698984b`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/698984b1d1b96c9b6ffaff00c7fc2e78e140e842) | [`r003.md`](./training-history/r003.md) |
 | r002 | Pollock 1.0 | [`5e571cd`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/5e571cde09ef2d1124e13d1c31d0b75f8bd0e513) | [`r002.md`](./training-history/r002.md) |
@@ -211,11 +211,11 @@ Use the Transformers example in the Polish section. The artifact uses standard `
 
 ### Fixed inference samples
 
-The shared [`fixed-sampling-v1`](./inference-samples/README.md) suite runs the same four prompts on r001-r005. Published revisions are loaded from full immutable commit SHAs; the pre-publication r005 artifact is loaded locally. Every weight-file SHA-256 is verified before inference. The reference protocol uses macOS 26.2 on arm64, CPU float32 with one thread, Transformers 5.15.1, seed 1337 reset for every prompt, temperature 0.7, top-k 50, and a 100-new-token limit.
+The shared [`fixed-sampling-v1`](./inference-samples/README.md) suite runs the same four prompts on r001-r005. All revisions are loaded from full immutable commit SHAs, and every weight-file SHA-256 is verified before inference. The reference protocol uses macOS 26.2 on arm64, CPU float32 with one thread, Transformers 5.15.1, seed 1337 reset for every prompt, temperature 0.7, top-k 50, and a 100-new-token limit.
 
 | Revision | Release | Weights | History |
 |---|---|---|---|
-| **r005** | **Pollock 1.3 (review package)** | local artifact; commit pending | [`r005.md`](./training-history/r005.md) |
+| **r005** | **Pollock 1.3** | [`e780025`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/e780025f15e06bc3765a74d973906eb8a11c022c) | [`r005.md`](./training-history/r005.md) |
 | r004 | Pollock 1.2 | [`30feb81`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/30feb81e5097eef13b939e1827d98e0458bf602d) | [`r004.md`](./training-history/r004.md) |
 | r003 | Pollock 1.1 | [`698984b`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/698984b1d1b96c9b6ffaff00c7fc2e78e140e842) | [`r003.md`](./training-history/r003.md) |
 | r002 | Pollock 1.0 | [`5e571cd`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/5e571cde09ef2d1124e13d1c31d0b75f8bd0e513) | [`r002.md`](./training-history/r002.md) |

inference-samples/README.md

@@ -22,7 +22,7 @@ Exact token replay was verified in the recorded environment. Sampling may diverg
 | r002 | Pollock 1.0 | [`5e571cd`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/5e571cde09ef2d1124e13d1c31d0b75f8bd0e513) | `2aada5b26abe9ce3b70af393d77be4ed1a799a7205b9c0dd5cfaff8bcaabce68` |
 | r003 | Pollock 1.1 | [`698984b`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/698984b1d1b96c9b6ffaff00c7fc2e78e140e842) | `3a9964a0ba5f6d20ef6061d2a7e4442c1b001209ebe80f7b7148a25573decaa9` |
 | r004 | Pollock 1.2 | [`30feb81`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/30feb81e5097eef13b939e1827d98e0458bf602d) | `3c453f22d4bb70228e5cca79e183a425e0f0221e54bd782d12027c42c943c880` |
-| r005 | Pollock 1.3 | local review artifact (`UNPUBLISHED`) | `3db43dfa44622e8156f060fee6f16875e108db4f5f949969cb91c0ca8cdce32c` |
+| r005 | Pollock 1.3 | [`e780025`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/e780025f15e06bc3765a74d973906eb8a11c022c) | `3db43dfa44622e8156f060fee6f16875e108db4f5f949969cb91c0ca8cdce32c` |
 
 ## Results
 
@@ -356,7 +356,7 @@ Output SHA-256: `5767898a26ba2104eb5ede9f29195d39715573f09d9d415f58881e5000cd688
 
 ## Reproduce or verify
 
-Run from this directory. The check downloads approximately 1.9 GB of published model artifacts if they are not already cached; the unpublished r005 artifact is loaded from the package root.
+Run from this directory. The check downloads approximately 2.4 GB of published model artifacts if they are not already cached.
 
 ```bash
 python -m venv .venv

inference-samples/config.json

@@ -52,8 +52,7 @@
     {
       "revision_id": "r005",
       "release": "Pollock 1.3",
-      "commit": "UNPUBLISHED",
-      "local_path": "..",
+      "commit": "e780025f15e06bc3765a74d973906eb8a11c022c",
       "model_sha256": "3db43dfa44622e8156f060fee6f16875e108db4f5f949969cb91c0ca8cdce32c"
     }
   ],

inference-samples/generate.py

@@ -257,13 +257,8 @@ def render_markdown(config: dict[str, Any], results: dict[str, Any]) -> str:
     ]
     for revision in results["revisions"]:
         commit = revision["commit"]
-        if commit == "UNPUBLISHED":
-            weights = "local review artifact (`UNPUBLISHED`)"
-        else:
-            commit_url = (
-                f"https://huggingface.co/{results['model_id']}/commit/{commit}"
-            )
-            weights = f"[`{commit[:7]}`]({commit_url})"
+        commit_url = f"https://huggingface.co/{results['model_id']}/commit/{commit}"
+        weights = f"[`{commit[:7]}`]({commit_url})"
         lines.append(
             f"| {revision['revision_id']} | {revision['release']} | "
             f"{weights} | `{revision['model_sha256']}` |"
@@ -309,9 +304,8 @@ def render_markdown(config: dict[str, Any], results: dict[str, Any]) -> str:
         [
             "## Reproduce or verify",
             "",
-            "Run from this directory. The check downloads approximately 1.9 GB of "
-            "published model artifacts if they are not already cached; the unpublished "
-            "r005 artifact is loaded from the package root.",
+            "Run from this directory. The check downloads approximately 2.4 GB of "
+            "published model artifacts if they are not already cached.",
             "",
             "```bash",
             "python -m venv .venv",

inference-samples/results.json

@@ -1966,7 +1966,7 @@
     {
       "revision_id": "r005",
       "release": "Pollock 1.3",
-      "commit": "UNPUBLISHED",
+      "commit": "e780025f15e06bc3765a74d973906eb8a11c022c",
       "model_sha256": "3db43dfa44622e8156f060fee6f16875e108db4f5f949969cb91c0ca8cdce32c",
       "loaded_model_version": "1.3",
       "results": [

release_manifest.json

@@ -5,10 +5,9 @@
   "release": "Pollock 1.3",
   "model_id": "SlayerLab/pollock-mini-lm-125m",
   "publication": {
-    "weights_commit": null,
-    "version_tag": null,
-    "planned_version_tag": "v1.3",
-    "publication_commit_pending": true
+    "weights_commit": "e780025f15e06bc3765a74d973906eb8a11c022c",
+    "version_tag": "v1.3",
+    "publication_commit_pending": false
   },
   "source_checkpoint": {
     "path_in_training_workspace": "runs/pollock-r005-v1p3/checkpoints/ckpt-final.pt",
@@ -83,9 +82,9 @@
     "seed": 1337,
     "prompt_count": 4,
     "revision_count": 5,
-    "local_review_artifact": true,
+    "local_review_artifact": false,
     "exact_replay_verified": true,
-    "publication_commit_pending": true
+    "publication_commit_pending": false
   },
   "conversion": {
     "target_class": "GPT2LMHeadModel",
@@ -101,16 +100,16 @@
     ".gitattributes": {"sha256": "c759491a998899dbefaec4d51cc791e68c714a37dd9f9829020a368788fe3063"},
     "CHANGELOG.md": {"sha256": "74a87cefef26ff31a9b9828df875d8c3a68383e13e360ff7ce201cfceb9af915"},
     "LICENSE.md": {"sha256": "162b97898a956d43d0a5428d1999bac3e42e2a97eb01a79f73ad719a4a15c412"},
-    "README.md": {"sha256": "0388e3fc5f72d42a33314865e54e8deb8aec21a64d280631f887a94d213a74ad"},
+    "README.md": {"sha256": "1958e268e5ac6838941b8225d27f506bbce4e37c8f76a9045d5913129e874ad7"},
     "assets/pollock-mini-lm-avatar-320.png": {"sha256": "7be10cc9d0f4aedeb298d9a5d722b2b2ac5b2e219b3884916dbca16198f70750"},
     "benchmarks/english.json": {"sha256": "c9210b71334898140b543ce684e9209c1919bbcdceb43224f4c5770f4fb93db7"},
     "config.json": {"sha256": "ba6775a805995b19d738604bcb035fccf012a0f4596b156d9945fd3a55c50d0b"},
     "generation_config.json": {"sha256": "435beb27be51f0ed054f4a011e5109d125cdadc118b8799b18b155cc798d94d2"},
-    "inference-samples/README.md": {"sha256": "cdbde0e4532382d89f8a9c3c22e2dc6ee05de41630cec7954bff9ce985a97172"},
-    "inference-samples/config.json": {"sha256": "831d175dd1f29ca059f5cb8e72f9705f99d1b941cdaee1db348b5b3f3ac55891"},
-    "inference-samples/generate.py": {"sha256": "75592499eeda1eb0a1e7235b5a690ee9c4084fd1bbca8e706ad65d86e92e10e8"},
+    "inference-samples/README.md": {"sha256": "178c85ec0e003754371bde7aea0af55fe918c8d45f8cad79ec371264f52210f1"},
+    "inference-samples/config.json": {"sha256": "fb504c65e8c6e01aa8d7f235c7f2ab0a88b95b2acbd884ec1e1d0e083669cdec"},
+    "inference-samples/generate.py": {"sha256": "807a9946472bd0087dd2b9260ee90d1dea09785d24d4f923f028667c53a23933"},
     "inference-samples/requirements.txt": {"sha256": "584583335ffb3061aa62058aee725b0de25fed7523904fbd13fb4622f601943d"},
-    "inference-samples/results.json": {"sha256": "88ab361fd616679c32a5fec13584f2010572370026763fa3bd0808f04227157f"},
+    "inference-samples/results.json": {"sha256": "c2eee21056ddd95eb3ece7d5bcd93242e62d90302814c0178c6beb4425d81553"},
     "model.safetensors": {"sha256": "3db43dfa44622e8156f060fee6f16875e108db4f5f949969cb91c0ca8cdce32c"},
     "special_tokens_map.json": {"sha256": "8b2257a17ea997bb038f43b133aefec82344ad2b8abc2b8a02a6c0a994ed624e"},
     "tokenizer.json": {"sha256": "3733307577230bb4802d2c774d8e2323f7f64e4139c13712736a57daf91bdda1"},
@@ -119,6 +118,6 @@
     "training-history/r002.md": {"sha256": "2b777a46e34fad856f323742dd0adf7f2c8ba6ee73a2ba2106d727644b1c8ca8"},
     "training-history/r003.md": {"sha256": "9fb6a1e04bdadada6f8760cc64a96e538ee7d23b7fe112959d8ba45b89ca8ada"},
     "training-history/r004.md": {"sha256": "968191c3a314989e5bfc749cd79e105ae06d5622c24b0f870435e698c969ac63"},
-    "training-history/r005.md": {"sha256": "9933566e549db9fa0027e3e643f2b21eefdf2b79559236137a17ec8281e1835e"}
+    "training-history/r005.md": {"sha256": "a25c2932701695dd7402582510adbe71e50a6fcc2da0fb3c6224a846dfdb22df"}
   }
 }

training-history/r005.md

@@ -44,4 +44,4 @@ LAMBADA perplexity was **47.328614**. Benchmarks used complete splits, zero few-
 
 ## Conversion and release status
 
-The native checkpoint was converted to `GPT2LMHeadModel`. A deterministic `[2, 64]` parity probe produced maximum absolute logit error **4.76837158203125e-06**. The fixed inference suite was replayed exactly on its recorded macOS 26.2 arm64 environment against the local, hash-verified r005 artifact. This is a review package: the Hugging Face commit and version tag are intentionally pending, and nothing was pushed automatically.
+The native checkpoint was converted to `GPT2LMHeadModel`. A deterministic `[2, 64]` parity probe produced maximum absolute logit error **4.76837158203125e-06**. The fixed inference suite was replayed exactly on its recorded macOS 26.2 arm64 environment against immutable Hugging Face weights commit [`e780025`](https://huggingface.co/SlayerLab/pollock-mini-lm-125m/commit/e780025f15e06bc3765a74d973906eb8a11c022c), with the expected model-file SHA-256 verified before loading.