tt-x10-fsdp2-fa2 — step 117 (X10b training backend: FSDP2 vs Megatron)
GRPO checkpoint from the TaskTrove X10b backend experiment, trained on
Iris / CoreWeave (cw-rno2a,
4x8 H100)
. Base model
Qwen/Qwen3-Coder-30B-A3B-Instruct
, trained on
DCAgent/exp_rpt_multifile
with SkyRL + Terminus-2. Single variable vs the X8 longclip base:
trainer.strategy=fsdp2
(+ flash-attention-2, async). Campaign verifier is
pass_ratio
shaping.
Checkpoint selection
global_step_117
was selected by the
trailing-5 reward EMA
(alpha = 1/3) over the stitched
6-generation curve (steps 1-131, no gaps) — the highest-EMA checkpoint (EMA 0.2535; step reward 0.2793;
pass@8 0.3906; entropy 0.0215), inside the reward plateau. Export performed on-Iris from the sharded
FSDP checkpoint bank via the checkpoint-export job (
hf_upload_mode=latest
).
Run status — question ANSWERED; stopped at 131/400 (plateau)
X10b's question — does FSDP2 survive the X8 schedule that collapsed every Megatron arm — is answered
yes
: 131 stable steps with entropy falling 0.284 -> 0.0129 (~22x, exploration exhausted) and response
length held (11.9k -> 13.5k tokens; no length collapse). Reward plateaued from ~step 40
(20-step windows 40-131 flat within noise; peak raw 0.332 at step 116); the remaining horizon buys
nothing. Stopped by owner decision 2026-08-15 at logged step 131, banked 129.
See
training_logs/
for metrics.csv (stitched per-step curve + trailing-5 EMA), report.md,
reward_plot.png, and the launch rl_config.yaml. W&B (dogml/OpenThoughts-Agent): base
jjjhduya
,
final gen r7
0s0kinwg
.
Training Traces
penfever/tt-x10-fsdp2-fa2
—
1/8 systematic subsample
(every 8th trial of 89,425 across 5 run generations, uniform coverage of
steps 1-131). Full-set transfer from CoreWeave object storage was ~430 GB; the subsample is a
documented deviation, approved by the owner, for an arm whose question was answered before horizon.
Runs of laion tt-x10-fsdp2-fa2-117-30B on huggingface.co
227
Total runs
0
24-hour runs
-113
3-day runs
-92
7-day runs
94
30-day runs
More Information About tt-x10-fsdp2-fa2-117-30B huggingface.co Model
tt-x10-fsdp2-fa2-117-30B huggingface.co is an AI model on huggingface.co that provides tt-x10-fsdp2-fa2-117-30B's model effect (), which can be used instantly with this laion tt-x10-fsdp2-fa2-117-30B model. huggingface.co supports a free trial of the tt-x10-fsdp2-fa2-117-30B model, and also provides paid use of the tt-x10-fsdp2-fa2-117-30B. Support call tt-x10-fsdp2-fa2-117-30B model through api, including Node.js, Python, http.
tt-x10-fsdp2-fa2-117-30B huggingface.co is an online trial and call api platform, which integrates tt-x10-fsdp2-fa2-117-30B's modeling effects, including api services, and provides a free online trial of tt-x10-fsdp2-fa2-117-30B, you can try tt-x10-fsdp2-fa2-117-30B online for free by clicking the link below.
laion tt-x10-fsdp2-fa2-117-30B online free url in huggingface.co:
tt-x10-fsdp2-fa2-117-30B is an open source model from GitHub that offers a free installation service, and any user can find tt-x10-fsdp2-fa2-117-30B on GitHub to install. At the same time, huggingface.co provides the effect of tt-x10-fsdp2-fa2-117-30B install, users can directly use tt-x10-fsdp2-fa2-117-30B installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
tt-x10-fsdp2-fa2-117-30B install url in huggingface.co: