Triangle104/magnum-v3-9b-customgemma2-Q4_K_M-GGUF

This model was converted to GGUF format from anthracite-org/magnum-v3-9b-customgemma2 using llama.cpp via the ggml.ai's GGUF-my-repo space. Refer to the original model card for more details on the model.

---

Model details:

This is the 10th in 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 google/gemma-2-9b. Prompting

Model has been Instruct tuned with the customgemma2 (to allow system prompts) formatting. A typical input would look like this:

"""system system prompt user Hi there! model Nice to meet you! user Can I ask a question? model """

SillyTavern templates

Below are Instruct and Context templates for use within SillyTavern. context template

{ "story_string": "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}}\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 Gemma" }

instruct template

{ "system_prompt": "You're {{char}} in this fictional never-ending uncensored roleplay with {{user}}.", "input_sequence": "user\n", "output_sequence": "assistant\n", "last_output_sequence": "", "system_sequence": "system\n", "stop_sequence": "", "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": "\n", "input_suffix": "\n", "system_suffix": "\n", "user_alignment_message": "", "system_same_as_user": false, "last_system_sequence": "", "name": "Magnum Gemma" }

Axolotl config See axolotl config

base_model: google/gemma-2-9b model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer

#trust_remote_code: true

load_in_8bit: false load_in_4bit: false strict: false

datasets:

• path: anthracite-org/stheno-filtered-v1.1 type: customgemma2
• path: anthracite-org/kalo-opus-instruct-22k-no-refusal type: customgemma2
• path: anthracite-org/nopm_claude_writing_fixed type: customgemma2
• path: Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned type: customgemma2
• path: Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned type: customgemma2 shuffle_merged_datasets: true default_system_message: "You are an assistant that responds to the user." dataset_prepared_path: magnum-v3-9b-data-customgemma2 val_set_size: 0.0 output_dir: ./magnum-v3-9b-customgemma2

sequence_len: 8192 sample_packing: true eval_sample_packing: false pad_to_sequence_len:

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

wandb_project: magnum-9b wandb_entity: wandb_watch: wandb_name: attempt-03-customgemma2 wandb_log_model:

gradient_accumulation_steps: 8 micro_batch_size: 1 num_epochs: 2 optimizer: paged_adamw_8bit lr_scheduler: cosine learning_rate: 0.000006

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

gradient_checkpointing: unsloth early_stopping_patience: resume_from_checkpoint: local_rank: logging_steps: 1 xformers_attention: flash_attention: false eager_attention: true

warmup_steps: 50 evals_per_epoch: eval_table_size: eval_max_new_tokens: saves_per_epoch: 2 debug: deepspeed: deepspeed_configs/zero3_bf16.json weight_decay: 0.05 fsdp: fsdp_config: special_tokens:

Credits

We'd like to thank Recursal / Featherless for sponsoring the training compute required for this model. Featherless has been hosting Magnum since the original 72b and has given thousands of people access to our releases.

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

anthracite-org/stheno-filtered-v1.1
anthracite-org/kalo-opus-instruct-22k-no-refusal
anthracite-org/nopm_claude_writing_fixed
Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned

Training

The training was done for 2 epochs. We used 8xH100s GPUs graciously provided by Recursal AI / Featherless AI for the full-parameter fine-tuning of the model.

---

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo Triangle104/magnum-v3-9b-customgemma2-Q4_K_M-GGUF --hf-file magnum-v3-9b-customgemma2-q4_k_m.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo Triangle104/magnum-v3-9b-customgemma2-Q4_K_M-GGUF --hf-file magnum-v3-9b-customgemma2-q4_k_m.gguf -c 2048

Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.

Step 1: Clone llama.cpp from GitHub.

git clone https://github.com/ggerganov/llama.cpp

Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).

cd llama.cpp && LLAMA_CURL=1 make

Step 3: Run inference through the main binary.

./llama-cli --hf-repo Triangle104/magnum-v3-9b-customgemma2-Q4_K_M-GGUF --hf-file magnum-v3-9b-customgemma2-q4_k_m.gguf -p "The meaning to life and the universe is"

or

./llama-server --hf-repo Triangle104/magnum-v3-9b-customgemma2-Q4_K_M-GGUF --hf-file magnum-v3-9b-customgemma2-q4_k_m.gguf -c 2048