Text Generation
Transformers
Safetensors
qwen3_moe
programming
code generation
code
codeqwen
Mixture of Experts
coding
coder
qwen2
chat
qwen
qwen-coder
Qwen3-30B-A3B-Instruct-2507
Qwen3-30B-A3B
mixture of experts
128 experts
8 active experts
256k context
qwen3
finetune
brainstorm 20x
brainstorm
optional thinking
conversational
Update README.md
Browse files
README.md
CHANGED
|
@@ -71,6 +71,7 @@ For coding, programming set expert to:
|
|
| 71 |
- 6-8 for general work.
|
| 72 |
- 10 for moderate work.
|
| 73 |
- 12-16 for complex work, long projects, complex coding.
|
|
|
|
| 74 |
- And for longer context, and/or multi-turn -> increase experts by 1-2 to help with longer context/multi turn understanding.
|
| 75 |
|
| 76 |
Recommended settings - general:
|
|
@@ -78,12 +79,14 @@ Recommended settings - general:
|
|
| 78 |
- Temp .3 to .6 (+- .2)
|
| 79 |
- Topk of 20, 40 or 100
|
| 80 |
- Topp of .95 / min p of .05
|
|
|
|
| 81 |
- System prompt (optional) to focus the model better.
|
| 82 |
|
| 83 |
Creative Use Cases:
|
| 84 |
- Rep pen of 1.05 or higher, especially if using a lower quant / lower temps.
|
| 85 |
- Also use rep pen of 1.05 or higher with very short prompts.
|
| 86 |
- You can set active experts as low as "4" for creative use cases.
|
|
|
|
| 87 |
|
| 88 |
This is the refined version -V1.4- from this project (see this repo for all settings, details, system prompts, example generations etc etc):
|
| 89 |
|
|
|
|
| 71 |
- 6-8 for general work.
|
| 72 |
- 10 for moderate work.
|
| 73 |
- 12-16 for complex work, long projects, complex coding.
|
| 74 |
+
- Suggest min context window 4k to 8k.
|
| 75 |
- And for longer context, and/or multi-turn -> increase experts by 1-2 to help with longer context/multi turn understanding.
|
| 76 |
|
| 77 |
Recommended settings - general:
|
|
|
|
| 79 |
- Temp .3 to .6 (+- .2)
|
| 80 |
- Topk of 20, 40 or 100
|
| 81 |
- Topp of .95 / min p of .05
|
| 82 |
+
- Suggest min context window 4k to 8k.
|
| 83 |
- System prompt (optional) to focus the model better.
|
| 84 |
|
| 85 |
Creative Use Cases:
|
| 86 |
- Rep pen of 1.05 or higher, especially if using a lower quant / lower temps.
|
| 87 |
- Also use rep pen of 1.05 or higher with very short prompts.
|
| 88 |
- You can set active experts as low as "4" for creative use cases.
|
| 89 |
+
- Suggest min context window 4k to 8k.
|
| 90 |
|
| 91 |
This is the refined version -V1.4- from this project (see this repo for all settings, details, system prompts, example generations etc etc):
|
| 92 |
|