Backup Qwen3.5-397B-data custom SFT adapter qwen35-397b-data-qwen35-9b-custom-sft-20260630T042600Z

README.md

@@ -1,245 +1,207 @@
 ---
-library_name: peft
 base_model: Qwen/Qwen3.5-9B
+library_name: peft
 pipeline_tag: text-generation
-datasets:
-- benchflow/env0-experiment-trajectories
 tags:
 - base_model:adapter:Qwen/Qwen3.5-9B
 - lora
-- peft
-- sft
-- env-0
-- openhands
-- daytona
-- qwen
+- transformers
 ---
 
-# BenchFlow Qwen3.5-9B Env-0 Mobile SFT LoRA Adapter
-
-This repository now points to the current env-0-mobile PR828 SFT adapter:
-
-`env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z`
-
-It replaces the earlier SFT adapter that was documented for the Prime
-`general-agent` reproduction. The repository contains a PEFT LoRA adapter only;
-it does not contain the `Qwen/Qwen3.5-9B` base weights.
-
-## Release Summary
-
-| Field | Value |
-| --- | --- |
-| Adapter repo | `benchflow/benchflow-qwen35-9b` |
-| Current version | `env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z` |
-| Model tag | `env0-mobile-pr828-20260625` |
-| Base checkpoint | `Qwen/Qwen3.5-9B` |
-| Base checkpoint form | Full, non-quantized source checkpoint; frozen during LoRA SFT |
-| Adapter type | LoRA / PEFT |
-| Main training run | `env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z` |
-| Training task source | `env-0-mobile/tasks-eval` |
-| Training artifact source | `benchflow/env0-experiment-trajectories/experiments/env0-mobile-pr828/training/env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z` |
-| Teacher trajectory run | `pr828-env0-mobile-full300-azure-openhands-daytona-20260625T041236Z` |
-| Baseline run | `pr828-env0-mobile-baseline-full300-qwen35-9b-a100-20260625T054948Z` |
-| Post-SFT eval run | `pr828-env0-mobile-sft-full300-qwen35-9b-h100-20260625T100226Z` |
-| W&B run | `https://wandb.ai/benchflow-ai/env0-mobile-pr828-qwen35-sft-20260625-h100/runs/env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z` |
-| Published to main | `2026-06-25` |
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
 
-## Intended Use
+## Model Card Authors [optional]
 
-This adapter is an experiment artifact for measuring whether SFT on
-BenchFlow/Daytona/OpenHands env-0-mobile trajectories improves task pass rate
-for `Qwen/Qwen3.5-9B`. It is intended for controlled evaluation and further
-research, not for production autonomous operation.
+[More Information Needed]
 
-## Data Recipe
+## Model Card Contact
 
-The training rows were generated by running all 300 tasks under
-`env-0-mobile/tasks-eval` with:
+[More Information Needed]
+### Framework versions
 
-- BenchFlow PR `benchflow-ai/benchflow#828`;
-- Daytona sandboxes;
-- OpenHands ACP agent;
-- Azure GPT-5.4-mini teacher;
-- `bench train convert` to Prime-RL SFT-compatible JSONL.
-
-The canonical teacher dataset has:
-
-| Field | Value |
-| --- | --- |
-| Canonical rows | `300` |
-| Teacher pass count | `83/300` |
-| Source LLM exchanges | `2163` |
-| Rows with tool calls | `175` |
-| Skipped rows after canonicalization | `0` |
-
-Training data artifact:
-
-`https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/env0-mobile-pr828/pr828-env0-mobile-full300-azure-openhands-daytona-20260625T041236Z`
-
-## Training Parameters
-
-| Field | Value |
-| --- | --- |
-| Trainer | Custom Transformers + PEFT LoRA SFT |
-| Model loaded for SFT | `Qwen/Qwen3.5-9B` full BF16 base weights |
-| Quantization | None |
-| Adapter | LoRA |
-| LoRA rank | `32` |
-| LoRA alpha | `64` |
-| LoRA dropout | `0.05` |
-| Target modules | `q_proj`, `k_proj`, `v_proj`, `o_proj`, `gate_proj`, `up_proj`, `down_proj` |
-| Sequence length | `8192` |
-| Micro batch size | `1` |
-| Gradient accumulation | `8` |
-| Learning rate | `1e-4` |
-| Max steps | `300` |
-| Saved checkpoints | `100`, `200`, `300` |
-| Hardware | Prime Intellect 1x H100 80GB, `massedcompute`, `$2.35/hr` |
-| W&B project | `env0-mobile-pr828-qwen35-sft-20260625-h100` |
-
-The A100 40GB feasibility check failed with CUDA OOM at `max_length=8192`.
-The H100 run completed the full epoch successfully.
-
-## Training Result
-
-| Metric | Value |
-| --- | ---: |
-| Completed step | `300` |
-| Best step | `300` |
-| Best eval loss | `0.4590291380882263` |
-| Training rows | `300` |
-| Eval rows used during training | `1` |
-| Final adapter file | `adapter_model.safetensors` |
-| Final adapter size | `232818064` bytes |
-
-Training artifacts:
-
-`https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/env0-mobile-pr828/training/env0-mobile-pr828-qwen35-bf16-lora-8k-h100-20260625T084605Z`
-
-## Evaluation Results
-
-All rows below use the same 300-task `env-0-mobile/tasks-eval` denominator and
-canonicalized result selection.
-
-| Model / stage | Pass | Pass rate |
-| --- | ---: | ---: |
-| Azure GPT-5.4-mini teacher | `83/300` | `27.67%` |
-| Qwen3.5-9B base, self-hosted official full weights | `4/300` | `1.33%` |
-| Qwen3.5-9B SFT adapter | `16/300` | `5.33%` |
-
-Lift over base:
-
-- absolute: `+12` passes, `+4.00` percentage points;
-- relative pass-count lift: `4.0x`.
-
-On the subset of 83 tasks passed by the GPT-5.4-mini teacher:
-
-| Model / stage | Pass | Pass rate |
-| --- | ---: | ---: |
-| Qwen3.5-9B base | `3/83` | `3.61%` |
-| Qwen3.5-9B SFT adapter | `13/83` | `15.66%` |
-
-Post-SFT eval artifacts:
-
-`https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/env0-mobile-pr828/pr828-env0-mobile-sft-full300-qwen35-9b-h100-20260625T100226Z`
-
-Baseline artifacts:
-
-`https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/env0-mobile-pr828/pr828-env0-mobile-baseline-full300-qwen35-9b-a100-20260625T054948Z`
-
-## Fireworks-Hosted env-0 Standard60 Results
-
-A same-host Fireworks comparison has now been completed on the current
-`env-0/tasks` standard60 denominator. This is separate from the 300-task
-`env-0-mobile/tasks-eval` denominator above.
-
-| Field | Value |
-| --- | --- |
-| HF trajectory index | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35` |
-| Baseline deployment | `openai/accounts/bingran-you/deployments/env0-qwen3p5-9b-standard60` |
-| SFT deployment | `openai/accounts/bingran-you/deployments/benchflow-qwen35-9b-env0-mobile-sft-live` |
-| Agent / sandbox | OpenHands + Daytona |
-| Skill mode | `with-skill` |
-| Usage tracking | `off` |
-| SFT tool support | `supportsTools=true`; direct smoke returned structured OpenAI `message.tool_calls` |
-| GitHub experiment log | `https://github.com/benchflow-ai/env-0-experiment/blob/main/experiments/fireworks-qwen35-180/EXPERIMENT_LOG.md` |
-| GitHub experiment report | `https://github.com/benchflow-ai/env-0-experiment/blob/main/experiments/fireworks-qwen35-180/EXPERIMENT_REPORT.md` |
-| Detailed GitHub result doc | `https://github.com/benchflow-ai/env-0-experiment/blob/main/experiments/fireworks-qwen35-180/reports/2026-06-27-fireworks-qwen35-standard60-3trials.md` |
-
-### Final 3-trial comparison
-
-| Model / stage | Trials | Rows | Strict pass | Pass rate | Rows with tools | Unscored rows |
-| --- | ---: | ---: | ---: | ---: | ---: | ---: |
-| Fireworks Qwen3.5 baseline | 3 | 180 | `19/180` | `10.56%` | `175/180` | `0` |
-| Fireworks BenchFlow Qwen3.5 SFT | 3 | 180 | `24/180` | `13.33%` | `180/180` | `0` |
-
-SFT is directionally higher by `+5` passes, or `+2.78` percentage points.
-This should not be presented as a statistically proven pass-rate lift: a rough
-two-proportion check gives `z ~= 0.81`, two-sided `p ~= 0.42`, and a 95%
-confidence interval for the pass-rate delta of about `[-3.92 pp, +9.47 pp]`.
-
-### Trial breakdown
-
-| Model | Trial | Strict pass | HF folder |
-| --- | --- | ---: | --- |
-| Baseline | `trial-01-20260626T165651Z` | `5/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/baselines/qwen3p5-9b-standard60-3trials/trial-01-20260626T165651Z` |
-| Baseline | `trial-02-20260627T055735Z` | `7/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/baselines/qwen3p5-9b-standard60-3trials/trial-02-20260627T055735Z` |
-| Baseline | `trial-03-20260627T072748Z-proxy` | `7/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/baselines/qwen3p5-9b-standard60-3trials/trial-03-20260627T072748Z-proxy` |
-| SFT | `trial-01-20260627T022446Z` | `8/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/sft/benchflow-qwen35-9b-env0-mobile-sft-live-standard60-3trials/trial-01-20260627T022446Z` |
-| SFT | `trial-02-20260627T082221Z` | `6/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/sft/benchflow-qwen35-9b-env0-mobile-sft-live-standard60-3trials/trial-02-20260627T082221Z` |
-| SFT | `trial-03-20260627T090839Z` | `10/60` | `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/sft/benchflow-qwen35-9b-env0-mobile-sft-live-standard60-3trials/trial-03-20260627T090839Z` |
-
-### Aggregate artifacts
-
-- Baseline 180-row trajectories:
-  `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/baselines/qwen3p5-9b-standard60-3trials`
-- Baseline aggregate:
-  `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/baselines/qwen3p5-9b-standard60-3trials/aggregate`
-- SFT 180-row trajectories:
-  `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/sft/benchflow-qwen35-9b-env0-mobile-sft-live-standard60-3trials`
-- SFT aggregate:
-  `https://huggingface.co/datasets/benchflow/env0-experiment-trajectories/tree/main/experiments/fireworks-qwen35/sft/benchflow-qwen35-9b-env0-mobile-sft-live-standard60-3trials/aggregate`
-
-### Serving interpretation
-
-- The Fireworks SFT deployment/tool-call bridge is validated. The failed merged
-  model path that emitted textual `<tool_call>` blocks is not the deployment
-  path used for these final runs.
-- The SFT final rows all have nonzero OpenHands tool calls (`180/180`).
-- Baseline trial 03 used a local OpenAI-compatible proxy that forwarded to the
-  Fireworks baseline deployment and normalized OpenHands shell-tool aliases;
-  the upstream model host remained Fireworks.
-- Retry rows replace only unscored provider, sandbox, or verifier artifacts.
-  Scored model timeouts and scored task failures are retained.
-
-## Loading
-
-```python
-import torch
-from peft import PeftModel
-from transformers import AutoModelForCausalLM, AutoTokenizer
-
-base = AutoModelForCausalLM.from_pretrained(
-    "Qwen/Qwen3.5-9B",
-    torch_dtype=torch.bfloat16,
-    trust_remote_code=True,
-)
-model = PeftModel.from_pretrained(base, "benchflow/benchflow-qwen35-9b")
-tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-9B", trust_remote_code=True)
-```
-
-## Caveats
-
-- This is a LoRA adapter. Load it on top of `Qwen/Qwen3.5-9B`.
-- The 300-task env-0-mobile results and the 60-task env-0 standard60 results
-  use different denominators and should not be mixed in one pass-rate claim.
-- The Fireworks standard60 SFT run is directionally higher than the Fireworks
-  standard60 baseline, but the observed lift is not statistically strong at
-  three trials per task.
-- The strongest 300-task env-0-mobile lift came from auth-revoke and a small
-  number of gcal tasks; gdoc/gdrive/gmail/multi-invite remain weak and should
-  be analyzed before a second epoch.
-- The env-0-mobile Dockerfiles referenced
-  `ghcr.io/benchflow-ai/env-0-base:latest`, which was unavailable during the
-  run. The experiment used the public mirror
-  `ghcr.io/oliver-dowhiz/env-0-base:latest`.
+- PEFT 0.19.1

adapter_config.json

@@ -19,27 +19,29 @@
   "lora_alpha": 64,
   "lora_bias": false,
   "lora_dropout": 0.05,
+  "lora_ga_config": null,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "peft_version": "0.18.0",
+  "peft_version": "0.19.1",
   "qalora_group_size": 16,
   "r": 32,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "v_proj",
     "k_proj",
-    "q_proj",
     "up_proj",
+    "down_proj",
+    "v_proj",
+    "q_proj",
     "gate_proj",
-    "down_proj"
+    "o_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,
+  "use_bdlora": null,
   "use_dora": false,
   "use_qalora": false,
   "use_rslora": false

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb01c366d71c056bfc133717e7e782c1e4b67c3452cb2adde51b59c305031852
+oid sha256:eaa2b1a8de1c51548807d1c5781be45cc589c8150f8bebfd593a8d3f109ee358
 size 232818064