darkmaniac7 commited on
Commit
8c9c454
·
verified ·
1 Parent(s): b45f612

Card: TokForge links, credits and licences for every upstream model

Browse files

Adds the standard TokForge block and a Credits section naming every upstream model, creator and conversion source with its license; fixes metadata and prose.

Files changed (1) hide show
  1. README.md +16 -3
README.md CHANGED
@@ -11,6 +11,7 @@ tags:
11
  - qwen3
12
  - draft-model
13
  - text-generation
 
14
  pipeline_tag: text-generation
15
  inference: false
16
  base_model: Qwen/Qwen3-0.6B
@@ -19,6 +20,8 @@ base_model_relation: finetune
19
 
20
  # Qwen3-0.6B-kl-baseline-20k-MNN
21
 
 
 
22
  Small `Qwen3-0.6B` draft model exported for **TokForge + MNN** speculative decoding on Android.
23
 
24
  This is a practical, mobile-oriented draft bundle rather than a standard Transformers checkpoint. It is intended to be paired with larger Qwen3 targets inside TokForge.
@@ -41,7 +44,7 @@ This model is one of the strongest lightweight draft candidates we have for mobi
41
 
42
  ## Benchmark notes
43
 
44
- We do not publish per-device numbers for this experimental variant. Uplift varies by device, target model, and workload. For measured results on the current default draft line, see [TokForge-AccelerationPack-Draft](https://huggingface.co/darkmaniac7/TokForge-AccelerationPack-Draft).
45
 
46
  Training acceptance (`alpha`) at the final logged epoch:
47
 
@@ -107,7 +110,17 @@ Known-good draft-side config:
107
 
108
  ## TokForge
109
 
110
- - Website: [tokforge.ai](https://tokforge.ai)
111
- - Discord: [Join the Discord](https://discord.gg/Acv3CBtfVm)
 
 
 
 
112
 
113
  If you benchmark this on your own device, feel free to share results in Discord.
 
 
 
 
 
 
 
11
  - qwen3
12
  - draft-model
13
  - text-generation
14
+ - on-device
15
  pipeline_tag: text-generation
16
  inference: false
17
  base_model: Qwen/Qwen3-0.6B
 
20
 
21
  # Qwen3-0.6B-kl-baseline-20k-MNN
22
 
23
+ Published by [TokForge](https://tokforge.ai) for on-device inference.
24
+
25
  Small `Qwen3-0.6B` draft model exported for **TokForge + MNN** speculative decoding on Android.
26
 
27
  This is a practical, mobile-oriented draft bundle rather than a standard Transformers checkpoint. It is intended to be paired with larger Qwen3 targets inside TokForge.
 
44
 
45
  ## Benchmark notes
46
 
47
+ We do not publish per-device numbers for this experimental variant. Results vary by device, target model, and workload. For measured results on the current default draft line, see [TokForge-AccelerationPack-Draft](https://huggingface.co/darkmaniac7/TokForge-AccelerationPack-Draft).
48
 
49
  Training acceptance (`alpha`) at the final logged epoch:
50
 
 
110
 
111
  ## TokForge
112
 
113
+ - **Website:** https://tokforge.ai
114
+ - **Discord:** https://discord.gg/Acv3CBtfVm
115
+ - **Google Play:** https://play.google.com/store/apps/details?id=dev.tokforge
116
+ - **iOS TestFlight:** https://testflight.apple.com/join/jnufjzRr
117
+
118
+ Runs on-device in the TokForge app.
119
 
120
  If you benchmark this on your own device, feel free to share results in Discord.
121
+
122
+ ## Credits
123
+
124
+ - Qwen3-0.6B by the Qwen team (Alibaba Cloud) ([Hugging Face](https://huggingface.co/Qwen/Qwen3-0.6B)). Used under the Apache-2.0 license.
125
+ - Qwen3-8B by the Qwen team (Alibaba Cloud) ([Hugging Face](https://huggingface.co/Qwen/Qwen3-8B)). Used under the Apache-2.0 license. Used as the distillation teacher.
126
+ - Converted with llmexport from [MNN](https://github.com/alibaba/MNN) by Alibaba (Apache-2.0).