config.json

{
  "_name_or_path": "/export/work/sboughorbel/bloomz/",
  "apply_residual_connection_post_layernorm": false,
  "architectures": [
    "BloomForCausalLM"
  ],
  "attention_dropout": 0.0,
  "attention_softmax_in_fp32": true,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "hidden_dropout": 0.0,
  "hidden_size": 14336,
  "initializer_range": 0.02,
  "layer_norm_epsilon": 1e-05,
  "masked_softmax_fusion": true,
  "model_type": "bloom",
  "n_head": 112,
  "n_layer": 70,
  "pad_token_id": 3,
  "pretraining_tp": 4,
  "quantization_config": {
    "llm_int8_enable_fp32_cpu_offload": false,
    "llm_int8_skip_modules": null,
    "llm_int8_threshold": 6.0,
    "load_in_8bit": true
  },
  "seq_length": 2048,
  "slow_but_exact": false,
  "torch_dtype": "float16",
  "transformers_version": "4.28.0",
  "use_cache": true,
  "vocab_size": 250880
}