ahpah commited on
Commit
63e6d5b
·
verified ·
1 Parent(s): 39fe4e3

Document verified Ollama generation for the five-epoch release

Browse files
Files changed (2) hide show
  1. README.md +3 -1
  2. ollama_smoke_test_5epochs.json +11 -0
README.md CHANGED
@@ -16,7 +16,7 @@ language:
16
 
17
  # YokoCoder-4B
18
 
19
- YokoCoder-4B is a compact, locally runnable coding assistant built on the abliterated Qwen3.5 4B model. It was fine-tuned for coding, further trained on Flappy Bird examples, and continued for five more epochs on a new coding dataset. Tested on a Flappy Bird generation task, it delivers good results for its size.
20
 
21
  ## Model details
22
 
@@ -45,3 +45,5 @@ YokoCoder-4B is a fine-tuned and quantized derivative of [Qwen3.5-4B-abliterated
45
  ## Latest training update
46
 
47
  The October 29, 2026 update continues the existing YokoCoder checkpoint for five epochs, completing 20 optimizer steps with training loss 1.0559. The Q4_K_M file contains the merged updated weights. The continued-training LoRA adapter and its training records are included under `adapters/dataset4-5epochs/`. Export hashes and provenance are in `export_manifest_5epochs.json`.
 
 
 
16
 
17
  # YokoCoder-4B
18
 
19
+ YokoCoder-4B is a compact, locally runnable coding assistant built on the abliterated Qwen3.5 4B model. It was fine-tuned for coding, further trained on Flappy Bird examples, and continued for five more epochs on a new coding dataset. It builds on the author's checkpoint that passed the Flappy Bird game-generation test. This five-epoch release was tested in Ollama and passed a three-case Python function smoke test, offering practical coding value for its size.
20
 
21
  ## Model details
22
 
 
45
  ## Latest training update
46
 
47
  The October 29, 2026 update continues the existing YokoCoder checkpoint for five epochs, completing 20 optimizer steps with training loss 1.0559. The Q4_K_M file contains the merged updated weights. The continued-training LoRA adapter and its training records are included under `adapters/dataset4-5epochs/`. Export hashes and provenance are in `export_manifest_5epochs.json`.
48
+
49
+ The local generation check and model identity are recorded in [ollama_smoke_test_5epochs.json](ollama_smoke_test_5epochs.json).
ollama_smoke_test_5epochs.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "yokocoder-4b",
3
+ "gguf_sha256": "a9292f20ceb1704004c0a27350dc44536c1dd337b960befa926ad81b2515cc77",
4
+ "architecture": "qwen35",
5
+ "parameters": "num_ctx 16384\nnum_predict -1\ntemperature 0.6\ntop_p 0.95",
6
+ "response": "```python\ndef add(a, b):\n return a + b\n```",
7
+ "generated_tokens": 18,
8
+ "done_reason": "stop",
9
+ "code_smoke_test": "passed",
10
+ "cases": 3
11
+ }