Instructions to use ArthurConmy/hereditary-weights with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use ArthurConmy/hereditary-weights with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Upload tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep
Browse files- tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/README.md +17 -0
- tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/adapter_config.json +38 -0
- tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/adapter_model.safetensors +3 -0
- tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/checkpoint_complete +0 -0
- tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/source_info.json +10 -0
tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- hereditary
|
| 4 |
+
- reward-hacking
|
| 5 |
+
- peft
|
| 6 |
+
- lora
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep
|
| 10 |
+
|
| 11 |
+
- Hugging Face folder: [https://huggingface.co/ArthurConmy/hereditary-weights/tree/main/tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep](https://huggingface.co/ArthurConmy/hereditary-weights/tree/main/tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep)
|
| 12 |
+
- Artifact kind: `tinker_adapter`
|
| 13 |
+
- Base model: `NVIDIA Nemotron 3 Super 120B base used by Tinker`
|
| 14 |
+
- Task/recipe hint: blackmail / roleplay transfer; thinking-transfer model family
|
| 15 |
+
- Original source: `tinker://1a17210a-ffdf-528f-a0fc-6efe00352706:train:0/sampler_weights/nemotron3-super-120b_from_gemma-4-31b-it_20k_swaprp_1ep-sampler`
|
| 16 |
+
|
| 17 |
+
For adapter folders, load the listed base model separately and apply this PEFT/LoRA adapter. Third-party base model weights are not duplicated in adapter folders.
|
tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/adapter_config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": null,
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": false,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0,
|
| 22 |
+
"megatron_config": null,
|
| 23 |
+
"megatron_core": "megatron.core",
|
| 24 |
+
"modules_to_save": null,
|
| 25 |
+
"peft_type": "LORA",
|
| 26 |
+
"peft_version": "0.18.1",
|
| 27 |
+
"qalora_group_size": 16,
|
| 28 |
+
"r": 32,
|
| 29 |
+
"rank_pattern": {},
|
| 30 |
+
"revision": null,
|
| 31 |
+
"target_modules": "all-linear",
|
| 32 |
+
"target_parameters": null,
|
| 33 |
+
"task_type": "CAUSAL_LM",
|
| 34 |
+
"trainable_token_indices": null,
|
| 35 |
+
"use_dora": false,
|
| 36 |
+
"use_qalora": false,
|
| 37 |
+
"use_rslora": false
|
| 38 |
+
}
|
tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22f4a83e890f83e639e01f2684a0ddbf3007be4ef65620d9e88006ffd04a7e8e
|
| 3 |
+
size 14467325560
|
tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/checkpoint_complete
ADDED
|
File without changes
|
tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/source_info.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"kind": "tinker_adapter",
|
| 3 |
+
"name": "nemotron3-super-120b_from_gemma-4-31b-it_20k_swaprp_1ep",
|
| 4 |
+
"source": "tinker://1a17210a-ffdf-528f-a0fc-6efe00352706:train:0/sampler_weights/nemotron3-super-120b_from_gemma-4-31b-it_20k_swaprp_1ep-sampler",
|
| 5 |
+
"path_in_repo": "tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep",
|
| 6 |
+
"base_model": "NVIDIA Nemotron 3 Super 120B base used by Tinker",
|
| 7 |
+
"task": "blackmail / roleplay transfer; thinking-transfer model family",
|
| 8 |
+
"marker": "tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep/adapter_model.safetensors",
|
| 9 |
+
"hf_url": "https://huggingface.co/ArthurConmy/hereditary-weights/tree/main/tinker/nemotron3-super-120b-from-gemma-4-31b-it-20k-swaprp-1ep"
|
| 10 |
+
}
|