Spaces:

Shasidharyadavr
/

netweaver_sre

Running

App Files Files Community

22.4 MB

Ctrl+K

Ctrl+K

5 contributors

History: 189 commits

Shasidharyadavr

docs(blog): auto-regenerate portfolio table from latest run

32f1c68 3 months ago

docs
docs: add training runbook capturing both Run #1 (heuristic) and Run #2 (HF Jobs GRPO) steps 3 months ago
logs
v3: bug fixes + composed rewards + real training evidence 3 months ago
notebooks
Update notebooks/train_unsloth_netweaver.ipynb 3 months ago
scratch
feat(grpo): launch v20..v26 mega-batch + auto-regen BLOG infrastructure 3 months ago
scripts
fix(grpo): support pre-quantized model mirrors; relaunch v10/v11/v12 as v16/v17/v18 3 months ago
server
GRPO 3B v25 (temp=0.7, LR=5e-05, r=32, model=granite-3.1-2b-instruct): 0.734->0.808 delta=+0.074 3 months ago
.gitattributes
1.58 kB
feat: enhance playground visuals and align with Phase 2 validation spec 4 months ago
.gitignore
447 Bytes
docs(grpo): post-mortem with 4-recipe ablation and root cause 3 months ago
BLOG.md
26.6 kB
docs(blog): auto-regenerate portfolio table from latest run 3 months ago
Dockerfile
806 Bytes
feat: enhance playground visuals and align with Phase 2 validation spec 4 months ago
README.md
47 kB
docs(submission): add BLOG.md mini-blog + README submission-materials section 3 months ago
__init__.py
454 Bytes
Add missing file: __init__.py 4 months ago
analytics_metrics.txt
5.53 kB
feat: enhance playground visuals and align with Phase 2 validation spec 4 months ago
client.py
1.95 kB
Initial NetWeaver-SRE OpenEnv submission 4 months ago
graders.py
8.73 kB
refactor: integrate OpenEnv Rubric system into grading logic and update dependencies 3 months ago
inference.py
9.87 kB
v3: bug fixes + composed rewards + real training evidence 3 months ago
leaderboard_results.json
29 kB
feat: add zero-shot LLM leaderboard + 3B GRPO job script 3 months ago
local_validator.py
6.15 kB
fix: clamp all scores strictly to (0.001, 0.999) - fix Phase 2 validation 4 months ago
models.py
3.92 kB
Finalize SRE scoring engine Models and anti-hacking tests 4 months ago
openenv.yaml
10.3 kB
v3: bug fixes + composed rewards + real training evidence 3 months ago
pyproject.toml
1.46 kB
refactor: integrate OpenEnv Rubric system into grading logic and update dependencies 3 months ago
reward_shaper.py
5.85 kB
v3: bug fixes + composed rewards + real training evidence 3 months ago
rubrics.py
13.2 kB
refactor: integrate OpenEnv Rubric system into grading logic and update dependencies 3 months ago
start.sh
204 Bytes
feat: enhance playground visuals and align with Phase 2 validation spec 4 months ago
test_endpoints.py
3.5 kB
fix: clamp all scores strictly to (0.001, 0.999) - fix Phase 2 validation 4 months ago
test_hard_live.py
3.64 kB
Optimize Hard Mode inference strategy and add transcript saving 4 months ago
test_logic.py
2.78 kB
feat: enhance playground visuals and align with Phase 2 validation spec 4 months ago
test_reward_hacking.py
7.71 kB
Finalize SRE scoring engine Models and anti-hacking tests 4 months ago
train_grpo.py
27.4 kB
feat(grpo): final portfolio results + small-model winning narrative + v19 3 months ago
training_results.json
1.83 kB
final: real heuristic training data + /metrics + /healthz endpoints 3 months ago
training_results_grpo_3b.json
15.9 kB
feat(v4): GRPO 3B real run + UI fixes + bug fixes + leaderboard polish 3 months ago
training_results_grpo_3b_v13.json
1.79 kB
GRPO 3B v13 (temp=0.7, LR=2e-05, r=32, model=zephyr-7b-beta): 0.729->0.623 delta=-0.107 3 months ago
training_results_grpo_3b_v14.json
2.1 kB
GRPO 3B v14 (temp=0.7, LR=8e-05, r=32, model=SmolLM2-1.7B-Instruct): 0.701->0.927 delta=+0.226 3 months ago
training_results_grpo_3b_v15.json
1.8 kB
GRPO 3B v15 (temp=0.7, LR=2e-05, r=32, model=granite-3.1-8b-instruct): 0.593->0.634 delta=+0.041 3 months ago
training_results_grpo_3b_v16.json
2.01 kB
GRPO 3B v16 (temp=0.7, LR=5e-05, r=64, model=Llama-3.2-3B-Instruct-bnb-4bit): 0.708->0.640 delta=-0.068 3 months ago
training_results_grpo_3b_v17.json
1.84 kB
GRPO 3B v17 (temp=0.7, LR=2e-05, r=32, model=Meta-Llama-3.1-8B-Instruct-bnb-4bit): 0.638->0.544 delta=-0.094 3 months ago
training_results_grpo_3b_v18.json
1.77 kB
GRPO 3B v18 (temp=0.7, LR=2e-05, r=32, model=gemma-2-9b-it-bnb-4bit): 0.691->0.946 delta=+0.254 3 months ago
training_results_grpo_3b_v2.json
1.59 kB
GRPO 3B v2 (calibrated, temp=0.7, LR=3e-05, r=32): 0.870->0.320 delta=-0.550 3 months ago
training_results_grpo_3b_v22.json
2 kB
GRPO 3B v22 (temp=0.7, LR=5e-05, r=32, model=gemma-2-2b-it-bnb-4bit): 0.596->0.658 delta=+0.062 3 months ago
training_results_grpo_3b_v24.json
1.74 kB
GRPO 3B v24 (temp=0.7, LR=2e-05, r=32, model=Qwen2.5-Math-7B-Instruct): 0.573->0.534 delta=-0.039 3 months ago
training_results_grpo_3b_v25.json
2 kB
GRPO 3B v25 (temp=0.7, LR=5e-05, r=32, model=granite-3.1-2b-instruct): 0.734->0.808 delta=+0.074 3 months ago
training_results_grpo_3b_v26.json
1.85 kB
GRPO 3B v26 (temp=0.7, LR=2e-05, r=32, model=Qwen2.5-Coder-7B-Instruct): 0.882->0.893 delta=+0.010 3 months ago
training_results_grpo_3b_v3.json
1.87 kB
GRPO 3B v3 (temp=0.8, LR=5e-05, r=64, model=Qwen2.5-3B-Instruct): 0.826->0.320 delta=-0.506 3 months ago
training_results_grpo_3b_v4.json
1.95 kB
GRPO 3B v4 (temp=0.8, LR=0.0001, r=64, model=Qwen2.5-1.5B-Instruct): 0.622->0.320 delta=-0.302 3 months ago
training_results_grpo_3b_v5.json
2.01 kB
GRPO 3B v5 (temp=0.8, LR=5e-05, r=64, model=Qwen2.5-3B-Instruct): 0.821->0.879 delta=+0.059 3 months ago
training_results_grpo_3b_v6.json
1.88 kB
GRPO 3B v6 (temp=0.8, LR=0.0001, r=16, model=Qwen2.5-0.5B-Instruct): 0.678->0.658 delta=-0.020 3 months ago
training_results_grpo_3b_v8.json
2.48 kB
GRPO 3B v8 (temp=0.8, LR=5e-05, r=64, model=Qwen2.5-3B-Instruct): 0.608->0.447 delta=-0.161 3 months ago
training_results_grpo_3b_v9.json
1.68 kB
GRPO 3B v9 (temp=0.7, LR=2e-05, r=32, model=Qwen2.5-7B-Instruct): 0.999->0.742 delta=-0.257 3 months ago
uv.lock
576 kB
Add uv.lock for multi-mode deployment validation 4 months ago