Monibee-Fudgekins commited on
Commit
b3352e6
·
verified ·
1 Parent(s): c01f2ff

Update model card (full run, pass@1 100.00%)

Browse files
Files changed (1) hide show
  1. README.md +77 -14
README.md CHANGED
@@ -1,21 +1,84 @@
1
  ---
2
- base_model: Monibee-Fudgekins/gemma-coder
 
 
 
 
 
3
  tags:
4
- - text-generation-inference
5
- - transformers
6
- - unsloth
7
- - gemma4
8
- license: apache-2.0
9
- language:
10
- - en
 
 
 
 
 
 
 
 
 
 
11
  ---
12
 
13
- # Uploaded finetuned model
14
 
15
- - **Developed by:** Monibee-Fudgekins
16
- - **License:** apache-2.0
17
- - **Finetuned from model :** Monibee-Fudgekins/gemma-coder
 
18
 
19
- This gemma4 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
 
 
20
 
21
- [<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: gemma
3
+ base_model: google/gemma-4-26B-A4B-it
4
+ datasets:
5
+ - ise-uiuc/Magicoder-OSS-Instruct-75K
6
+ library_name: transformers
7
+ pipeline_tag: text-generation
8
  tags:
9
+ - code
10
+ - coding-assistant
11
+ - qlora
12
+ - unsloth
13
+ model-index:
14
+ - name: gemma-coder
15
+ results:
16
+ - task:
17
+ type: text-generation
18
+ name: Code generation
19
+ dataset:
20
+ type: code-eval
21
+ name: remote-agent-dev-platform coding_eval
22
+ metrics:
23
+ - type: pass@1
24
+ value: 1.0000
25
+ name: pass@1 (Python/JS/React/Go/Java/Swift)
26
  ---
27
 
28
+ # gemma-coder
29
 
30
+ Coding-focused fine-tune of [`google/gemma-4-26B-A4B-it`](https://huggingface.co/google/gemma-4-26B-A4B-it)
31
+ (**Gemma 4 26B A4B**, an MoE with ~4B active params), produced automatically by the
32
+ weekly retrain pipeline in [remote-agent-dev-platform](https://github.com/Monibee-Fudgekins/remote-agent-dev-platform).
33
+ **Last updated: 2026-07-22 16:45 UTC** · run mode: `full` · promoted: **True**.
34
 
35
+ ## Model description
36
+ QLoRA fine-tune of google/gemma-4-26B-A4B-it specialized for coding assistance. It is the
37
+ default agent model for the remote-agent-dev-platform (served via vLLM on Modal).
38
 
39
+ ## Intended uses & limitations
40
+ - **Intended:** code generation and assistance in Python, JavaScript/React, Go, Java,
41
+ and Swift, inside a sandboxed agent that runs/tests the output.
42
+ - **Not intended:** safety-critical use, or running generated code unreviewed.
43
+ - **Limitations:** a small, free-tier-trained model — it can produce incorrect or
44
+ insecure code. Always review and test. Quality tracks the training data, which is
45
+ still being built out.
46
+
47
+ ## Training data
48
+ - Dataset: [`ise-uiuc/Magicoder-OSS-Instruct-75K`](https://huggingface.co/datasets/ise-uiuc/Magicoder-OSS-Instruct-75K)
49
+
50
+ ## Training procedure
51
+ - Method: QLoRA (Unsloth), 4-bit base, LoRA r=8 / alpha=16 on
52
+ attention only, lr 2e-05,
53
+ max seq len 768, optimizer adamw_8bit.
54
+ - Progress: **cycle 1 — 2047 / 4000 steps** (trained
55
+ in weekly ~8h chunks on Kaggle's free 2×T4, resuming each week; training is
56
+ continuous — a finished cycle rolls into the next).
57
+
58
+ ## Evaluation
59
+ Sandboxed multi-language **pass@1** harness (`finetune/evaluate.py`): the model
60
+ completes functions that are then compiled/run against unit tests. Languages whose
61
+ toolchain is unavailable are skipped.
62
+
63
+ **Overall pass@1: 100.00%** over 11 executed problems
64
+ (0 skipped). Promotion threshold: 80%.
65
+
66
+ | language | passed / run | pass@1 |
67
+ |---|---|---|
68
+ | javascript | 3/3 | 100.00% |
69
+ | python | 8/8 | 100.00% |
70
+
71
+ ## How to use
72
+ ```python
73
+ from transformers import AutoModelForCausalLM, AutoTokenizer
74
+ tok = AutoTokenizer.from_pretrained("Monibee-Fudgekins/gemma-coder")
75
+ model = AutoModelForCausalLM.from_pretrained("Monibee-Fudgekins/gemma-coder", device_map="auto")
76
+ msgs = [{"role": "user", "content": "Write a Python function that reverses a string."}]
77
+ ids = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
78
+ print(tok.decode(model.generate(ids, max_new_tokens=256)[0]))
79
+ ```
80
+
81
+
82
+ ## Provenance
83
+ Generated by `finetune/kaggle/run.py` in [https://github.com/Monibee-Fudgekins/remote-agent-dev-platform](https://github.com/Monibee-Fudgekins/remote-agent-dev-platform); see that repo for
84
+ the full training + eval pipeline.