Instructions to use MoritzMo123/kfac-svd_uniform_mistral-7b_0.9 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MoritzMo123/kfac-svd_uniform_mistral-7b_0.9 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MoritzMo123/kfac-svd_uniform_mistral-7b_0.9", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("MoritzMo123/kfac-svd_uniform_mistral-7b_0.9", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MoritzMo123/kfac-svd_uniform_mistral-7b_0.9 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MoritzMo123/kfac-svd_uniform_mistral-7b_0.9

SGLang

How to use MoritzMo123/kfac-svd_uniform_mistral-7b_0.9 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MoritzMo123/kfac-svd_uniform_mistral-7b_0.9",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MoritzMo123/kfac-svd_uniform_mistral-7b_0.9 with Docker Model Runner:
```
docker model run hf.co/MoritzMo123/kfac-svd_uniform_mistral-7b_0.9
```

kfac-svd_uniform_mistral-7b_0.9 / config.json

MoritzMo123

Upload folder using huggingface_hub

022a64f verified 3 months ago

Raw

History Blame Contribute Delete

12.5 kB

	{
	"auto_map": {
	"AutoConfig": "configuration_low_rank.LowRankConfig",
	"AutoModelForCausalLM": "modeling_low_rank.LowRankCausalLM"
	},
	"base_config": {
	"_name_or_path": "unsloth/mistral-7b",
	"add_cross_attention": false,
	"architectures": [
	"MistralForCausalLM"
	],
	"attention_dropout": 0.0,
	"bad_words_ids": null,
	"begin_suppress_tokens": null,
	"bos_token_id": 1,
	"chunk_size_feed_forward": 0,
	"cross_attention_hidden_size": null,
	"decoder_start_token_id": null,
	"diversity_penalty": 0.0,
	"do_sample": false,
	"dtype": "float16",
	"early_stopping": false,
	"encoder_no_repeat_ngram_size": 0,
	"eos_token_id": 2,
	"exponential_decay_length_penalty": null,
	"finetuning_task": null,
	"forced_bos_token_id": null,
	"forced_eos_token_id": null,
	"head_dim": 128,
	"hidden_act": "silu",
	"hidden_size": 4096,
	"id2label": {
	"0": "LABEL_0",
	"1": "LABEL_1"
	},
	"initializer_range": 0.02,
	"intermediate_size": 14336,
	"is_decoder": false,
	"is_encoder_decoder": false,
	"label2id": {
	"LABEL_0": 0,
	"LABEL_1": 1
	},
	"length_penalty": 1.0,
	"max_length": 20,
	"max_position_embeddings": 32768,
	"min_length": 0,
	"model_type": "mistral",
	"no_repeat_ngram_size": 0,
	"num_attention_heads": 32,
	"num_beam_groups": 1,
	"num_beams": 1,
	"num_hidden_layers": 32,
	"num_key_value_heads": 8,
	"num_return_sequences": 1,
	"output_attentions": false,
	"output_hidden_states": false,
	"output_scores": false,
	"pad_token_id": 0,
	"prefix": null,
	"problem_type": null,
	"pruned_heads": {},
	"remove_invalid_values": false,
	"repetition_penalty": 1.0,
	"return_dict": true,
	"return_dict_in_generate": false,
	"rms_norm_eps": 1e-05,
	"rope_theta": 10000.0,
	"sep_token_id": null,
	"sliding_window": 4096,
	"suppress_tokens": null,
	"task_specific_params": null,
	"temperature": 1.0,
	"tf_legacy_loss": false,
	"tie_encoder_decoder": false,
	"tie_word_embeddings": false,
	"tokenizer_class": null,
	"top_k": 50,
	"top_p": 1.0,
	"torch_dtype": "float16",
	"torchscript": false,
	"transformers_version": "4.55.0",
	"typical_p": 1.0,
	"unsloth_version": "2024.9",
	"use_bfloat16": false,
	"use_cache": true,
	"vocab_size": 32000
	},
	"base_model_id": "unsloth/mistral-7b",
	"base_model_type": "mistral",
	"compression_method": "kfac_svd",
	"dtype": "float16",
	"model_type": "low_rank_compressed",
	"rank_dict": {
	"model.layers.0.mlp.down_proj": 2866,
	"model.layers.0.mlp.gate_proj": 2866,
	"model.layers.0.mlp.up_proj": 2866,
	"model.layers.0.self_attn.k_proj": 737,
	"model.layers.0.self_attn.o_proj": 1843,
	"model.layers.0.self_attn.q_proj": 1843,
	"model.layers.0.self_attn.v_proj": 737,
	"model.layers.1.mlp.down_proj": 2866,
	"model.layers.1.mlp.gate_proj": 2866,
	"model.layers.1.mlp.up_proj": 2866,
	"model.layers.1.self_attn.k_proj": 737,
	"model.layers.1.self_attn.o_proj": 1843,
	"model.layers.1.self_attn.q_proj": 1843,
	"model.layers.1.self_attn.v_proj": 737,
	"model.layers.10.mlp.down_proj": 2866,
	"model.layers.10.mlp.gate_proj": 2866,
	"model.layers.10.mlp.up_proj": 2866,
	"model.layers.10.self_attn.k_proj": 737,
	"model.layers.10.self_attn.o_proj": 1843,
	"model.layers.10.self_attn.q_proj": 1843,
	"model.layers.10.self_attn.v_proj": 737,
	"model.layers.11.mlp.down_proj": 2866,
	"model.layers.11.mlp.gate_proj": 2866,
	"model.layers.11.mlp.up_proj": 2866,
	"model.layers.11.self_attn.k_proj": 737,
	"model.layers.11.self_attn.o_proj": 1843,
	"model.layers.11.self_attn.q_proj": 1843,
	"model.layers.11.self_attn.v_proj": 737,
	"model.layers.12.mlp.down_proj": 2866,
	"model.layers.12.mlp.gate_proj": 2866,
	"model.layers.12.mlp.up_proj": 2866,
	"model.layers.12.self_attn.k_proj": 737,
	"model.layers.12.self_attn.o_proj": 1843,
	"model.layers.12.self_attn.q_proj": 1843,
	"model.layers.12.self_attn.v_proj": 737,
	"model.layers.13.mlp.down_proj": 2866,
	"model.layers.13.mlp.gate_proj": 2866,
	"model.layers.13.mlp.up_proj": 2866,
	"model.layers.13.self_attn.k_proj": 737,
	"model.layers.13.self_attn.o_proj": 1843,
	"model.layers.13.self_attn.q_proj": 1843,
	"model.layers.13.self_attn.v_proj": 737,
	"model.layers.14.mlp.down_proj": 2866,
	"model.layers.14.mlp.gate_proj": 2866,
	"model.layers.14.mlp.up_proj": 2866,
	"model.layers.14.self_attn.k_proj": 737,
	"model.layers.14.self_attn.o_proj": 1843,
	"model.layers.14.self_attn.q_proj": 1843,
	"model.layers.14.self_attn.v_proj": 737,
	"model.layers.15.mlp.down_proj": 2866,
	"model.layers.15.mlp.gate_proj": 2866,
	"model.layers.15.mlp.up_proj": 2866,
	"model.layers.15.self_attn.k_proj": 737,
	"model.layers.15.self_attn.o_proj": 1843,
	"model.layers.15.self_attn.q_proj": 1843,
	"model.layers.15.self_attn.v_proj": 737,
	"model.layers.16.mlp.down_proj": 2866,
	"model.layers.16.mlp.gate_proj": 2866,
	"model.layers.16.mlp.up_proj": 2866,
	"model.layers.16.self_attn.k_proj": 737,
	"model.layers.16.self_attn.o_proj": 1843,
	"model.layers.16.self_attn.q_proj": 1843,
	"model.layers.16.self_attn.v_proj": 737,
	"model.layers.17.mlp.down_proj": 2866,
	"model.layers.17.mlp.gate_proj": 2866,
	"model.layers.17.mlp.up_proj": 2866,
	"model.layers.17.self_attn.k_proj": 737,
	"model.layers.17.self_attn.o_proj": 1843,
	"model.layers.17.self_attn.q_proj": 1843,
	"model.layers.17.self_attn.v_proj": 737,
	"model.layers.18.mlp.down_proj": 2866,
	"model.layers.18.mlp.gate_proj": 2866,
	"model.layers.18.mlp.up_proj": 2866,
	"model.layers.18.self_attn.k_proj": 737,
	"model.layers.18.self_attn.o_proj": 1843,
	"model.layers.18.self_attn.q_proj": 1843,
	"model.layers.18.self_attn.v_proj": 737,
	"model.layers.19.mlp.down_proj": 2866,
	"model.layers.19.mlp.gate_proj": 2866,
	"model.layers.19.mlp.up_proj": 2866,
	"model.layers.19.self_attn.k_proj": 737,
	"model.layers.19.self_attn.o_proj": 1843,
	"model.layers.19.self_attn.q_proj": 1843,
	"model.layers.19.self_attn.v_proj": 737,
	"model.layers.2.mlp.down_proj": 2866,
	"model.layers.2.mlp.gate_proj": 2866,
	"model.layers.2.mlp.up_proj": 2866,
	"model.layers.2.self_attn.k_proj": 737,
	"model.layers.2.self_attn.o_proj": 1843,
	"model.layers.2.self_attn.q_proj": 1843,
	"model.layers.2.self_attn.v_proj": 737,
	"model.layers.20.mlp.down_proj": 2866,
	"model.layers.20.mlp.gate_proj": 2866,
	"model.layers.20.mlp.up_proj": 2866,
	"model.layers.20.self_attn.k_proj": 737,
	"model.layers.20.self_attn.o_proj": 1843,
	"model.layers.20.self_attn.q_proj": 1843,
	"model.layers.20.self_attn.v_proj": 737,
	"model.layers.21.mlp.down_proj": 2866,
	"model.layers.21.mlp.gate_proj": 2866,
	"model.layers.21.mlp.up_proj": 2866,
	"model.layers.21.self_attn.k_proj": 737,
	"model.layers.21.self_attn.o_proj": 1843,
	"model.layers.21.self_attn.q_proj": 1843,
	"model.layers.21.self_attn.v_proj": 737,
	"model.layers.22.mlp.down_proj": 2866,
	"model.layers.22.mlp.gate_proj": 2866,
	"model.layers.22.mlp.up_proj": 2866,
	"model.layers.22.self_attn.k_proj": 737,
	"model.layers.22.self_attn.o_proj": 1843,
	"model.layers.22.self_attn.q_proj": 1843,
	"model.layers.22.self_attn.v_proj": 737,
	"model.layers.23.mlp.down_proj": 2866,
	"model.layers.23.mlp.gate_proj": 2866,
	"model.layers.23.mlp.up_proj": 2866,
	"model.layers.23.self_attn.k_proj": 737,
	"model.layers.23.self_attn.o_proj": 1843,
	"model.layers.23.self_attn.q_proj": 1843,
	"model.layers.23.self_attn.v_proj": 737,
	"model.layers.24.mlp.down_proj": 2866,
	"model.layers.24.mlp.gate_proj": 2866,
	"model.layers.24.mlp.up_proj": 2866,
	"model.layers.24.self_attn.k_proj": 737,
	"model.layers.24.self_attn.o_proj": 1843,
	"model.layers.24.self_attn.q_proj": 1843,
	"model.layers.24.self_attn.v_proj": 737,
	"model.layers.25.mlp.down_proj": 2866,
	"model.layers.25.mlp.gate_proj": 2866,
	"model.layers.25.mlp.up_proj": 2866,
	"model.layers.25.self_attn.k_proj": 737,
	"model.layers.25.self_attn.o_proj": 1843,
	"model.layers.25.self_attn.q_proj": 1843,
	"model.layers.25.self_attn.v_proj": 737,
	"model.layers.26.mlp.down_proj": 2866,
	"model.layers.26.mlp.gate_proj": 2866,
	"model.layers.26.mlp.up_proj": 2866,
	"model.layers.26.self_attn.k_proj": 737,
	"model.layers.26.self_attn.o_proj": 1843,
	"model.layers.26.self_attn.q_proj": 1843,
	"model.layers.26.self_attn.v_proj": 737,
	"model.layers.27.mlp.down_proj": 2866,
	"model.layers.27.mlp.gate_proj": 2866,
	"model.layers.27.mlp.up_proj": 2866,
	"model.layers.27.self_attn.k_proj": 737,
	"model.layers.27.self_attn.o_proj": 1843,
	"model.layers.27.self_attn.q_proj": 1843,
	"model.layers.27.self_attn.v_proj": 737,
	"model.layers.28.mlp.down_proj": 2866,
	"model.layers.28.mlp.gate_proj": 2866,
	"model.layers.28.mlp.up_proj": 2866,
	"model.layers.28.self_attn.k_proj": 737,
	"model.layers.28.self_attn.o_proj": 1843,
	"model.layers.28.self_attn.q_proj": 1843,
	"model.layers.28.self_attn.v_proj": 737,
	"model.layers.29.mlp.down_proj": 2866,
	"model.layers.29.mlp.gate_proj": 2866,
	"model.layers.29.mlp.up_proj": 2866,
	"model.layers.29.self_attn.k_proj": 737,
	"model.layers.29.self_attn.o_proj": 1843,
	"model.layers.29.self_attn.q_proj": 1843,
	"model.layers.29.self_attn.v_proj": 737,
	"model.layers.3.mlp.down_proj": 2866,
	"model.layers.3.mlp.gate_proj": 2866,
	"model.layers.3.mlp.up_proj": 2866,
	"model.layers.3.self_attn.k_proj": 737,
	"model.layers.3.self_attn.o_proj": 1843,
	"model.layers.3.self_attn.q_proj": 1843,
	"model.layers.3.self_attn.v_proj": 737,
	"model.layers.30.mlp.down_proj": 2866,
	"model.layers.30.mlp.gate_proj": 2866,
	"model.layers.30.mlp.up_proj": 2866,
	"model.layers.30.self_attn.k_proj": 737,
	"model.layers.30.self_attn.o_proj": 1843,
	"model.layers.30.self_attn.q_proj": 1843,
	"model.layers.30.self_attn.v_proj": 737,
	"model.layers.31.mlp.down_proj": 2866,
	"model.layers.31.mlp.gate_proj": 2866,
	"model.layers.31.mlp.up_proj": 2866,
	"model.layers.31.self_attn.k_proj": 737,
	"model.layers.31.self_attn.o_proj": 1843,
	"model.layers.31.self_attn.q_proj": 1843,
	"model.layers.31.self_attn.v_proj": 737,
	"model.layers.4.mlp.down_proj": 2866,
	"model.layers.4.mlp.gate_proj": 2866,
	"model.layers.4.mlp.up_proj": 2866,
	"model.layers.4.self_attn.k_proj": 737,
	"model.layers.4.self_attn.o_proj": 1843,
	"model.layers.4.self_attn.q_proj": 1843,
	"model.layers.4.self_attn.v_proj": 737,
	"model.layers.5.mlp.down_proj": 2866,
	"model.layers.5.mlp.gate_proj": 2866,
	"model.layers.5.mlp.up_proj": 2866,
	"model.layers.5.self_attn.k_proj": 737,
	"model.layers.5.self_attn.o_proj": 1843,
	"model.layers.5.self_attn.q_proj": 1843,
	"model.layers.5.self_attn.v_proj": 737,
	"model.layers.6.mlp.down_proj": 2866,
	"model.layers.6.mlp.gate_proj": 2866,
	"model.layers.6.mlp.up_proj": 2866,
	"model.layers.6.self_attn.k_proj": 737,
	"model.layers.6.self_attn.o_proj": 1843,
	"model.layers.6.self_attn.q_proj": 1843,
	"model.layers.6.self_attn.v_proj": 737,
	"model.layers.7.mlp.down_proj": 2866,
	"model.layers.7.mlp.gate_proj": 2866,
	"model.layers.7.mlp.up_proj": 2866,
	"model.layers.7.self_attn.k_proj": 737,
	"model.layers.7.self_attn.o_proj": 1843,
	"model.layers.7.self_attn.q_proj": 1843,
	"model.layers.7.self_attn.v_proj": 737,
	"model.layers.8.mlp.down_proj": 2866,
	"model.layers.8.mlp.gate_proj": 2866,
	"model.layers.8.mlp.up_proj": 2866,
	"model.layers.8.self_attn.k_proj": 737,
	"model.layers.8.self_attn.o_proj": 1843,
	"model.layers.8.self_attn.q_proj": 1843,
	"model.layers.8.self_attn.v_proj": 737,
	"model.layers.9.mlp.down_proj": 2866,
	"model.layers.9.mlp.gate_proj": 2866,
	"model.layers.9.mlp.up_proj": 2866,
	"model.layers.9.self_attn.k_proj": 737,
	"model.layers.9.self_attn.o_proj": 1843,
	"model.layers.9.self_attn.q_proj": 1843,
	"model.layers.9.self_attn.v_proj": 737
	},
	"target_ratio": 0.9,
	"torch_dtype": "float16",
	"transformers_version": "4.55.0"
	}