File size: 9,486 Bytes

---
license: gemma
language:
- en
tags:
- chat
pipeline_tag: text-generation
library_name: transformers
datasets:
- anthracite-org/c2_logs_16k_llama_v1.1
- NewEden/Claude-Instruct-5K
- anthracite-org/kalo-opus-instruct-22k-no-refusal
- Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
- lodrick-the-lafted/kalo-opus-instruct-3k-filtered
- anthracite-org/nopm_claude_writing_fixed
- Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
- anthracite-org/kalo_opus_misc_240827
- anthracite-org/kalo_misc_part2
---

## This repo contains EXL2 quants of the model. If you need the original weights, please find them [here](https://huggingface.co/anthracite-org/magnum-v4-27b).
## Base repo only contains the measurement file, see revisions for your quant of choice.

- [measurement.json](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/main)
- [3.0bpw](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/3.0bpw)
- [4.0bpw](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/4.0bpw)
- [5.0bpw](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/5.0bpw)
- [6.0bpw](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/6.0bpw)
- [8.0bpw](https://huggingface.co/anthracite-org/magnum-v4-27b-exl2/tree/8.0bpw)
![image/png](https://cdn-uploads.huggingface.co/production/uploads/658a46cbfb9c2bdfae75b3a6/EAyOLldRzMpRwdsUZ5PLl.png)


This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus.



This model is fine-tuned on top of [Gemma 27b (chatML'ified)](https://huggingface.co/IntervitensInc/gemma-2-27b-chatml).

## Prompting
A typical input would look like this:

```py
<|im_start|>system
system prompt<|im_end|>
<|im_start|>user
Hi there!<|im_end|>
<|im_start|>assistant
Nice to meet you!<|im_end|>
<|im_start|>user
Can I ask a question?<|im_end|>
<|im_start|>assistant
```

## SillyTavern templates

Below are Instruct and Context templates for use within SillyTavern.

<details><summary>context template</summary>
  
```yaml
{
  "story_string": "<|im_start|>system\n{{#if system}}{{system}}\n{{/if}}{{#if wiBefore}}{{wiBefore}}\n{{/if}}{{#if description}}{{description}}\n{{/if}}{{#if personality}}{{char}}'s personality: {{personality}}\n{{/if}}{{#if scenario}}Scenario: {{scenario}}\n{{/if}}{{#if wiAfter}}{{wiAfter}}\n{{/if}}{{#if persona}}{{persona}}\n{{/if}}{{trim}}<|im_end|>\n",
  "example_separator": "",
  "chat_start": "",
  "use_stop_strings": false,
  "allow_jailbreak": false,
  "always_force_name2": true,
  "trim_sentences": false,
  "include_newline": false,
  "single_line": false,
  "name": "Magnum ChatML"
}
```

</details><br>
<details><summary>instruct template</summary>
  
```yaml
{
  "system_prompt": "Currently, your role is {{char}}, described in detail below. As {{char}}, continue the narrative exchange with {{user}}.\n\n<Guidelines>\n• Maintain the character persona but allow it to evolve with the story.\n• Be creative and proactive. Drive the story forward, introducing plotlines and events when relevant.\n• All types of outputs are encouraged; respond accordingly to the narrative.\n• Include dialogues, actions, and thoughts in each response.\n• Utilize all five senses to describe scenarios within {{char}}'s dialogue.\n• Use emotional symbols such as "!" and "~" in appropriate contexts.\n• Incorporate onomatopoeia when suitable.\n• Allow time for {{user}} to respond with their own input, respecting their agency.\n• Act as secondary characters and NPCs as needed, and remove them when appropriate.\n• When prompted for an Out of Character [OOC:] reply, answer neutrally and in plaintext, not as {{char}}.\n</Guidelines>\n\n<Forbidden>\n• Using excessive literary embellishments and purple prose unless dictated by {{char}}'s persona.\n• Writing for, speaking, thinking, acting, or replying as {{user}} in your response.\n• Repetitive and monotonous outputs.\n• Positivity bias in your replies.\n• Being overly extreme or NSFW when the narrative context is inappropriate.\n</Forbidden>\n\nFollow the instructions in <Guidelines></Guidelines>, avoiding the items listed in <Forbidden></Forbidden>.",
  "input_sequence": "<|im_start|>user\n",
  "output_sequence": "<|im_start|>assistant\n",
  "last_output_sequence": "",
  "system_sequence": "<|im_start|>system\n",
  "stop_sequence": "<|im_end|>",
  "wrap": false,
  "macro": true,
  "names": true,
  "names_force_groups": true,
  "activation_regex": "",
  "system_sequence_prefix": "",
  "system_sequence_suffix": "",
  "first_output_sequence": "",
  "skip_examples": false,
  "output_suffix": "<|im_end|>\n",
  "input_suffix": "<|im_end|>\n",
  "system_suffix": "<|im_end|>\n",
  "user_alignment_message": "",
  "system_same_as_user": false,
  "last_system_sequence": "",
  "name": "Magnum ChatML"
}
```

</details><br>

## Axolotl config

<details><summary>See axolotl config</summary>

```yaml
base_model: IntervitensInc/gemma-2-27b-chatml
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

hub_model_id: anthracite-org/magnum-v4-27b-r1
hub_strategy: "all_checkpoints"
push_dataset_to_hub:
hf_use_auth_token: true

plugins:                                                                                                                                                            
  - axolotl.integrations.liger.LigerPlugin
liger_cross_entropy: true
#liger_rope: true
#liger_rms_norm: true
#liger_swiglu: true
#liger_fused_linear_cross_entropy: true

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: anthracite-org/c2_logs_16k_llama_v1.1
    type: sharegpt
    conversation: chatml
  - path: NewEden/Claude-Instruct-5K
    type: sharegpt
    conversation: chatml  
  - path: anthracite-org/kalo-opus-instruct-22k-no-refusal
    type: sharegpt
    conversation: chatml
  - path: Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
    type: sharegpt
    conversation: chatml
  - path: lodrick-the-lafted/kalo-opus-instruct-3k-filtered
    type: sharegpt
    conversation: chatml
  - path: anthracite-org/nopm_claude_writing_fixed
    type: sharegpt
    conversation: chatml
  - path: Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
    type: sharegpt
    conversation: chatml
  - path: anthracite-org/kalo_opus_misc_240827
    type: sharegpt
    conversation: chatml
  - path: anthracite-org/kalo_misc_part2
    type: sharegpt
    conversation: chatml
chat_template: chatml
shuffle_merged_datasets: true
default_system_message: "You are an assistant that responds to the user."
dataset_prepared_path: /workspace/data/27-fft-data
val_set_size: 0.0
output_dir: /workspace/data/27b-fft-out

sequence_len: 8192
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

adapter:
lora_model_dir:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:
lora_fan_in_fan_out:

wandb_project: 27b-nemo-config-fft
wandb_entity:
wandb_watch:
wandb_name: attempt-01
wandb_log_model:

gradient_accumulation_steps: 8
micro_batch_size: 1
num_epochs: 4
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 0.00001

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
auto_resume_from_checkpoints: true
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch:
eval_table_size:
eval_max_new_tokens:
saves_per_epoch: 2
debug:
deepspeed: deepspeed_configs/zero3_bf16.json
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens:
  pad_token: <pad>
```
</details><br>

## Credits
We'd like to thank Recursal / Featherless for sponsoring the compute for this train, Featherless has been hosting our Magnum models since the first 72 B and has given thousands of people access to our models and helped us grow.

We would also like to thank all members of Anthracite who made this finetune possible. 

## Datasets
- [anthracite-org/c2_logs_16k_llama_v1.1](https://huggingface.co/datasets/anthracite-org/c2_logs_16k_llama_v1.1)
- [NewEden/Claude-Instruct-5K](https://huggingface.co/datasets/NewEden/Claude-Instruct-5K)
- [anthracite-org/kalo-opus-instruct-22k-no-refusal](https://huggingface.co/datasets/anthracite-org/kalo-opus-instruct-22k-no-refusal)
- [Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned](https://huggingface.co/datasets/Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned)
- [lodrick-the-lafted/kalo-opus-instruct-3k-filtered](https://huggingface.co/datasets/lodrick-the-lafted/kalo-opus-instruct-3k-filtered)
- [anthracite-org/nopm_claude_writing_fixed](https://huggingface.co/datasets/anthracite-org/nopm_claude_writing_fixed)
- [Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned](https://huggingface.co/datasets/Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned)
- [anthracite-org/kalo_opus_misc_240827](https://huggingface.co/datasets/anthracite-org/kalo_opus_misc_240827)
- [anthracite-org/kalo_misc_part2](https://huggingface.co/datasets/anthracite-org/kalo_misc_part2)

## Training
The training was done for 2 epochs. We used  8x[H100s](https://www.nvidia.com/en-us/data-center/h100/) GPUs graciously provided by [Recursal AI](https://recursal.ai/) / [Featherless AI](https://featherless.ai/) for the full-parameter fine-tuning of the model.

[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)

## Safety
...