RichardErkhov/Menouar_-_gemma-2b-chat-ultra-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

gemma-2b-chat-ultra - GGUF

• —Model creator: https://huggingface.co/Menouar/
• —Original model: https://huggingface.co/Menouar/gemma-2b-chat-ultra/

Original model description:

license: other tags:

• —generatedfromtrainer
• —google/gemma
• —PyTorch
• —transformers
• —trl
• —peft
• —tensorboard model-index:
• —name: gemma-2b-chat-ultra results: [] datasets:
• —HuggingFaceH4/deita-10k-v0-sft licensename: gemma-terms-of-use licenselink: https://ai.google.dev/gemma/terms language:
• —en base_model: google/gemma-2b widget:
• —example_title: LLM messages:
• —role: user content: What is a large language model (LLM)? pipeline_tag: text-generation ---

Model Card for gemma-2b-chat-ultra:

💬🤖

gemma-2b-chat-ultra is a language model that is trained to act as AI assistant. It is a finetuned version of google/gemma-2b that was trained using SFTTrainer on publicly available dataset HuggingFaceH4/deita-10k-v0-sft.

Training Metrics

The training metrics can be found on **TensorBoard**.

Training hyperparameters

The following hyperparameters were used during the training:

• —output_dir: peft-lora-model

• —overwriteoutputdir: True

• —do_train: False

• —do_eval: False

• —do_predict: False

• —evaluation_strategy: no

• —predictionlossonly: False

• —perdevicetrainbatchsize: 2

• —perdeviceevalbatchsize: None

• —pergputrainbatchsize: None

• —pergpuevalbatchsize: None

• —gradientaccumulationsteps: 4

• —evalaccumulationsteps: None

• —eval_delay: 0

• —learning_rate: 2e-05

• —weight_decay: 0.0

• —adam_beta1: 0.9

• —adam_beta2: 0.999

• —adam_epsilon: 1e-08

• —maxgradnorm: 0.3

• —numtrainepochs: 3

• —max_steps: -1

• —lrschedulertype: cosine

• —lrschedulerkwargs: {}

• —warmup_ratio: 0.1

• —warmup_steps: 0

• —log_level: passive

• —loglevelreplica: warning

• —logoneach_node: True

• —loggingdir: peft-lora-model/runs/Mar2107-18-25_a518bf4a6403

• —logging_strategy: steps

• —loggingfirststep: False

• —logging_steps: 10

• —loggingnaninf_filter: True

• —save_strategy: epoch

• —save_steps: 500

• —savetotallimit: None

• —save_safetensors: True

• —saveoneach_node: False

• —saveonlymodel: False

• —no_cuda: False

• —use_cpu: False

• —usempsdevice: False

• —seed: 42

• —data_seed: None

• —jitmodeeval: False

• —use_ipex: False

• —bf16: True

• —fp16: False

• —fp16optlevel: O1

• —halfprecisionbackend: auto

• —bf16fulleval: False

• —fp16fulleval: False

• —tf32: None

• —local_rank: 0

• —ddp_backend: None

• —tpunumcores: None

• —tpumetricsdebug: False

• —debug: []

• —dataloaderdroplast: False

• —eval_steps: None

• —dataloadernumworkers: 0

• —dataloaderprefetchfactor: None

• —past_index: -1

• —run_name: peft-lora-model

• —disable_tqdm: False

• —removeunusedcolumns: True

• —label_names: None

• —loadbestmodelatend: False

• —metricforbest_model: None

• —greaterisbetter: None

• —ignoredataskip: False

• —fsdp: []

• —fsdpminnum_params: 0

• —fsdpconfig: {'minnumparams': 0, 'xla': False, 'xlafsdpv2': False, 'xlafsdpgradckpt': False}

• —fsdptransformerlayerclsto_wrap: None

• —acceleratorconfig: AcceleratorConfig(splitbatches=False, dispatchbatches=None, evenbatches=True, useseedablesampler=True)

• —deepspeed: None

• —labelsmoothingfactor: 0.0

• —optim: adamwtorchfused

• —optim_args: None

• —adafactor: False

• —groupbylength: False

• —lengthcolumnname: length

• —report_to: ['tensorboard']

• —ddpfindunused_parameters: None

• —ddpbucketcap_mb: None

• —ddpbroadcastbuffers: None

• —dataloaderpinmemory: True

• —dataloaderpersistentworkers: False

• —skipmemorymetrics: True

• —uselegacyprediction_loop: False

• —pushtohub: False

• —resumefromcheckpoint: None

• —hubmodelid: None

• —hubstrategy: everysave

• —hub_token: None

• —hubprivaterepo: False

• —hubalwayspush: False

• —gradient_checkpointing: True

• —gradientcheckpointingkwargs: {'use_reentrant': False}

• —includeinputsfor_metrics: False

• —fp16_backend: auto

• —pushtohubmodelid: None

• —pushtohub_organization: None

• —pushtohub_token: None

• —mp_parameters:

• —autofindbatch_size: False

• —full_determinism: False

• —torchdynamo: None

• —ray_scope: last

• —ddp_timeout: 1800

• —torch_compile: False

• —torchcompilebackend: None

• —torchcompilemode: None

• —dispatch_batches: None

• —split_batches: None

• —includetokensper_second: False

• —includenuminputtokensseen: False

• —neftunenoisealpha: None

• —distributed_state: Distributed environment: NO Num processes: 1 Process index: 0 Local process index: 0 Device: cuda

• —ngpu: 1

• —_cachedsetupdevices: cuda:0

• —deepspeed_plugin: None