alt.freeweights.text / stageC-pbs-80-8B
From: laion <huggingface.co/laion>Newsgroups: alt.freeweights.textSubject:

stageC-pbs-80-8B

Date: 2026-11-15Commit: 3159d85Signed: 2DC28CC7B26A283B ✓

Model with 8.2B parameters. Safetensors weights, 16.6 GB in all. Published by laion under Apache-2.0.

8.2B params16.6GBstageC-pbs-80-8Bsafetensorspermissive
FactsParams8.2BSize16.6 GBFormatssafetensorsTasknot taggedLicenseApache-2.0From sourcepermissive
$ fw get laion/stageC-pbs-80-8B
# manifest sha256 66ccc4ae0cad… · 16.6 GB · 46 files
$ fw verify ./stageC-pbs-80-8B laion/stageC-pbs-80-8B
From sourceHosted (signed)laion/stageC-pbs-80-8BLast check: 2026-11-15 05:34 UTC

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Fine-tune of GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink.

Works with

Runs with

Ranked by fit to this model (safetensors), then by GitHub stars.

Overview

Size
16.6 GB
Weights
16.4 GB (safetensors)
Parameters
8.2B
Formats
safetensors
Task
not tagged
License
Apache-2.0 (permissive)
Files
46, each with a sha256
Publisher
laion
Added
2026-11-15
Section
alt.freeweights.text
Lineage

Original release from the publisher.

Base model: laion/GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink

Where these bytes come from
Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources
BadgeHostLast checkResult
From sourcehuggingface.co2026-11-15 05:34 UTCpassed

Model card

From the publisher at commit 3159d85. Shown as plain text with links; images open at the publisher.

stageC-pbs-80-8B#

Agentic RL (SkyRL, FSDP2) checkpoint from the stageC cell of the a3 / pymethods2test agentic-RL family. This is the global_step_80 (max_steps) export — the legitimate end of training (trainer.max_steps=80).

  • Base model: laion/GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink (a Qwen3-8B SFT)
  • Dataset: DCAgent/exp_rpt_pymethods2test-large
  • Algorithm: RLOO-N (advantage_estimator=rloo_n_pbs), per-batch-shaped reward channel (enable_token_reward_channel=true), loss_reduction=seq_mean_token_sum_norm_global (seqnorm), TIS on (use_tis=true, tis_imp_ratio_cap=2.0), eps_clip=0.2/0.2, no KL loss.
  • Training: 14 nodes, train_batch_size=64, 2 epochs, max_steps=80, ckpt_interval=2, hf_save_interval=5. WANDB offline (Jupiter).
  • Sibling cells: laion/stageB-channel-80-8B, open-athena/stageD-thinkbudget-80-8B.

Final training metrics (global_step_80)#

metricvalue
reward (avg_raw)0.5645
pass@80.672
entropy0.290
raw_grad_norm~3.6e-5 (seqnorm global-denom artifact)
tis_imp_ratio_mean0.987
tis_imp_ratio_capped_fraction~1e-5

Training Traces#

Training-time Daytona/Harbor rollouts: open-athena/stageC-pbs (the last episode of each trial — the rollouts the policy trained on after rollback/truncation).

Contents#

4-shard safetensors weights + config + tokenizer/chat_template + generation_config + rl_config.json (the launch config) + training_logs/ (per-step metrics CSVs + vLLM metrics + raw .out chain logs — the W&B-equivalent, since Jupiter runs WANDB offline).

Files

46 files · 16.6 GBWeights 16.4 GB (safetensors). Sizes and sha256 from the signed manifest.
Show all 46 files with their sha256
FileSizesha256
.gitattributes2.15 KB5e6fd67ed6592bf340f5401d39fd70e5f6d085a94e76c0dd98b69d2c6de93acd
README.md1.91 KB1eefe1089af190ad56a1cdc9429dc0128c5a6cbcdc4802e71fe991ee7a0ab896
added_tokens.json707 Bc0284b582e14987fbd3d5a2cb2bd139084371ed9acbae488829a1c900833c680
chat_template.jinja4.17 KBa55ee1b1660128b7098723e0abcd92caa0788061051c62d51cbe87d9cf1974d8
config.json1.54 KBed36641e183b3d95dcefad8906aa22d183c445e4b582c8a0ca7d71a713c85f2d
generation_config.json188 B3b101cd89b140b757d0bca469e75570e75789a2423ada90a33ec17707dfacf08
merges.txt1.67 MB8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5
model-00001-of-00004.safetensors4.90 GBeb5bb9906e8f0f8b1d3dc56b5855f49918600a7a0beca8e746a46411610bfe9d
model-00002-of-00004.safetensors4.92 GBc673dac137318ef04fa01706821dfb26bfd8d27d1baa953429e6a355b13be4c8
model-00003-of-00004.safetensors4.98 GBa2a1876b480eea92d98966d5479d16277a47fc86153bcc570804c8f421b7efa2
model-00004-of-00004.safetensors1.58 GB1ad644d5effebfa0f7035c4bfc7963ae571efcf7958048a535d9255454cab78d
model.safetensors.index.json32.9 KBe7e4bdc58d3302c97357f27979b270987dd620cf0cd9b1c130e3a51c9d64df95
rl_config.json8.62 KBe2a6ab7615f6203a319af26296cc27912ad6fe2282aedbc68bfcce47a44809d5
special_tokens_map.json613 B76862e765266b85aa9459767e33cbaf13970f327a0e88d1c65846c2ddd3a1ecd
tokenizer.json11.4 MBaeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
tokenizer_config.json5.43 KB3e27b302e9844c5b9648a8dbc1958fbbf3bda640e07302f53c17ac7a3192def5
training_logs/20260618_160949_metrics_stageC-pbs_869079.csv6.13 KB4a936be4139c1024f61366d653aa3cdae55972292cf452d11fb7ad8301187507
training_logs/20260618_160949_metrics_stageC-pbs_869080.csv10.5 KBbb461fcbea53fb7042cbbb451c49fe63109fcb74f1a61794342142946d4b2c71
training_logs/20260618_160949_metrics_stageC-pbs_869081.csv10.8 KBf3bcf4d8fde82bf21b526636872506944cc364159381d7709231bd510ab0db53
training_logs/20260618_160949_metrics_stageC-pbs_869082.csv9.46 KB1fee8ac01d2f820dd493f3f81675ddfa08a6b731e12147ac00f50919a0e4967f
training_logs/20260618_160949_metrics_stageC-pbs_869083.csv10.4 KB0e1f6b220e57a0b9ce317d33ca2820000fd2224650cd98db953db6319ba5c4bc
training_logs/20260618_160949_metrics_stageC-pbs_869084.csv7.01 KB621fcae27bbdf0ffda612709348665adfb7a1998833bef81689519f000dc99ba
training_logs/20260618_160949_metrics_stageC-pbs_914214.csv3.68 KB6eecc5c69a849b27875bdfa9ba923f1024a8e37301270b9c0ad5cbc2dcf3db30
training_logs/20260618_160949_metrics_stageC-pbs_918535.csv4.13 KB38ea639612d57f161308f4d72488fe16d9a2f3ca01a92541e4daa7616a96dfe4
training_logs/20260618_160949_metrics_table.csv48.1 KB3f377b2918109654078c270f395af3624b0fbcc13a9fbe19efb82909fe1b61b5
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869079.csv363 KB693f93b5aaa590c827fb1b661a24a524dbb11e4b20514fe2fd8ef88c869b040f
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869080.csv639 KB77897ca786a3c8bbee2aae1433d5406bc45f68f17d920f636fdc7313da2e8506
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869081.csv646 KB9031acf352cbb5ca1d48e2f870affd5b4fdb8c28430785113b12bba8cdb47ac4
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869082.csv584 KBef8b8ed7c67a496d50a91c8f5d537ad7b4d029761319e6ffd852ad66da845fa0
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869083.csv648 KB9138da1a54b17d77106ff6d060c2312cbcfa7b67646a5d81916f4028c9e2379a
training_logs/20260618_160949_vllm_metrics_stageC-pbs_869084.csv271 KBc65c3ace030a7855ca519602580712c841f07af7dd15c930a430bc28015b84e9
training_logs/20260618_160949_vllm_metrics_stageC-pbs_914214.csv127 KBdac6b97b7b918190e77011fdda848dd9d7e0bacc41f6624d4e893a5a7cd11660
training_logs/20260618_160949_vllm_metrics_stageC-pbs_918535.csv154 KB9507ddc95a4cfd10c551969ecb41e79cfc8fbdcfd08a27a50f29b2adb61f48e2
training_logs/20260618_160949_vllm_metrics_table.csv4.24 MB3506bc358288b7a0cc7ff487f05108342b1b9505738ef33498a2415a2cf1a9f9
training_logs/stageC-pbs_869079.out20.6 MB0feff6711216575f19848c8fd1a337f573a93a5a110dc9bdc855f0e6f71335b0
training_logs/stageC-pbs_869080.out45.0 MBaea769f00d5be80e180baf13ed36836c5e4c112f2435ec19efec26c52dbaffb5
training_logs/stageC-pbs_869081.out41.5 MB21b20f6c6b3d2b82e0e4362eb4edf2fa9ddb35085e8e252b3f236b8a4ce56ccb
training_logs/stageC-pbs_869082.out40.6 MB1077513b7a5d90aa7a64c2f0dc8c9b697c4caff92641b0c12803cc48f83181de
training_logs/stageC-pbs_869083.out42.6 MBf043830672ab00864675123dd62a0889df5d01c34350ab64dfa8674d9058b182
training_logs/stageC-pbs_869084.out24.7 MB6c81c4c82638a689d41a4977dd6dd9673e910ac50b0bbea364a777d5c403f31e
training_logs/stageC-pbs_911577.out1.40 KBdd9250dcb81f830b0c8481fc6316548b0386c318bfd1b182b21cac7b3b60d14c
training_logs/stageC-pbs_911578.out1.40 KB05dc200dd69240bc1728c7bd880f46ac8922a768f509554743e5bc2cf168ebfb
training_logs/stageC-pbs_911579.out1.40 KBead55f49dfac97a4fc7c92c61c0757c75f23d457321dd78d00c0d1ffe573a5a3
training_logs/stageC-pbs_914214.out11.1 MBa2eeb8d6dfbb61fed5d3512b901a7ed701e71ff283894b21adfb01925aaf0464
training_logs/stageC-pbs_918535.out15.7 MBd9570b1ecb31e4812228261078c389aea376e19ba66ab5aab86ae213da7d1137
vocab.json2.78 MBca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910
VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.
Upstream commitDateContents
3159d8551f pinned now2026-11-15 05:34 UTC46 files · 16.6 GB

Use it

$ fw get laion/stageC-pbs-80-8B
# manifest sha256 66ccc4ae0cad… · 16.6 GB · 46 files
$ fw verify ./stageC-pbs-80-8B laion/stageC-pbs-80-8B
fw.pyChecks the signature and every sha256, tries each source in order, resumes.
$ python3 fw.py get https://freeweights.allternit.com/manifests/laion/stageC-pbs-80-8B.json --dir ./stageC-pbs-80-8B $ python3 fw.py verify ./stageC-pbs-80-8B ./laion__stageC-pbs-80-8B.json
hfThe Hugging Face CLI with one setting. Same repo id.
$ HF_ENDPOINT=https://freeweights.allternit.com hf download laion/stageC-pbs-80-8B --revision 3159d8551fcc94a213f1aa629167262c5b528b8f --local-dir ./stageC-pbs-80-8B

License

Name
Apache License 2.0
Id
Apache-2.0
Class
Permissive
Redistribution
yes
Listing
Hosted (signed)
Decided
License as declared by the publisher. No person read it. 2026-11-15
Text in
README.md
Text sha256
1eefe1089af190ad56a1cdc9429dc0128c5a6cbcdc4802e71fe991ee7a0ab896
Terms gate
No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: laion <huggingface.co/laion>Newsgroups: alt.freeweights.textSubject: stageC-pbs-80-8BDate: 2026-11-15Commit: 3159d85Signed: 2DC28CC7B26A283B ✓
Upstream
laion/stageC-pbs-80-8B
Pinned commit
3159d8551fcc94a213f1aa629167262c5b528b8f
Manifest
stageC-pbs-80-8B.json · .minisig
Manifest sha256
66ccc4ae0cadcfd07e27d8ddf68be369f206913209e4f4eb67bfb977c7707009
Signature
minisign, Ed25519, key 2DC28CC7B26A283B
Created
2026-11-15 05:34 UTC
Generator
freeweights-manifest/0.1.0
Takedowns
None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/laion/stageC-pbs-80-8B.json $ curl -O https://freeweights.allternit.com/manifests/laion/stageC-pbs-80-8B.json.minisig $ minisign -Vm stageC-pbs-80-8B.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq $ shasum -a 256 <file> # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

No model in the catalog names this one as its parent.

Get it
Download
$ HF_ENDPOINT=https://freeweights.allternit.com hf download laion/stageC-pbs-80-8B
Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest
minisign, key 2DC28CC7B26A283B. stageC-pbs-80-8B.json. Provenance. Takedown history: none.
From: laion <huggingface.co/laion>Subject: a2-rl-e2egit_large-50-32B
From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_e2egit-large-15-8B
From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_e2egit-v2-10-8B
From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_pymethods2test-large-80-8B
From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_exp_rpt_unitsyn-python-large-10-8B
From: laion <huggingface.co/laion>Subject: a3-rl-DCAgent_mix_h2_language_proportional-65-8B