K45 / config.json
Mhair's picture
Run 10. Outer Step 80. Inner Step 40.
2998896 verified
{
"_name_or_path": "distributed/llama-1b",
"all_reduce_scores": {
"0": "NOT_ALIVE",
"1": "NON_PARTICIPATING",
"10": "SUCCESS",
"100": "NOT_ALIVE",
"101": "SUCCESS",
"102": "SUCCESS",
"103": "NON_PARTICIPATING",
"104": "NOT_ALIVE",
"105": "SUCCESS",
"106": "NON_PARTICIPATING",
"107": "SUCCESS",
"108": "NON_PARTICIPATING",
"109": "SUCCESS",
"11": "NON_PARTICIPATING",
"110": "NON_PARTICIPATING",
"111": "SUCCESS",
"112": "NON_PARTICIPATING",
"113": "SUCCESS",
"114": "SUCCESS",
"115": "NON_PARTICIPATING",
"116": "NOT_ALIVE",
"117": "NON_PARTICIPATING",
"118": "NON_PARTICIPATING",
"119": "NON_PARTICIPATING",
"12": "NON_PARTICIPATING",
"120": "NOT_ALIVE",
"121": "SUCCESS",
"122": "NON_PARTICIPATING",
"123": "SUCCESS",
"124": "SUCCESS",
"125": "FAIL",
"126": "NON_PARTICIPATING",
"127": "NOT_ALIVE",
"128": "SUCCESS",
"129": "NON_PARTICIPATING",
"13": "SUCCESS",
"130": "NOT_ALIVE",
"131": "NOT_ALIVE",
"132": "NOT_ALIVE",
"133": "NOT_ALIVE",
"134": "NOT_ALIVE",
"135": "NON_PARTICIPATING",
"136": "NON_PARTICIPATING",
"137": "NOT_ALIVE",
"138": "NOT_ALIVE",
"139": "NOT_ALIVE",
"14": "NOT_ALIVE",
"140": "NON_PARTICIPATING",
"141": "NON_PARTICIPATING",
"142": "SUCCESS",
"143": "NOT_ALIVE",
"144": "NON_PARTICIPATING",
"145": "NOT_ALIVE",
"146": "SUCCESS",
"147": "NON_PARTICIPATING",
"148": "NON_PARTICIPATING",
"149": "NON_PARTICIPATING",
"15": "NON_PARTICIPATING",
"150": "NOT_ALIVE",
"151": "NOT_ALIVE",
"152": "NOT_ALIVE",
"153": "SUCCESS",
"154": "NON_PARTICIPATING",
"155": "NON_PARTICIPATING",
"156": "NON_PARTICIPATING",
"157": "NOT_ALIVE",
"158": "NON_PARTICIPATING",
"159": "SUCCESS",
"16": "NOT_ALIVE",
"160": "NOT_ALIVE",
"161": "SUCCESS",
"162": "NOT_ALIVE",
"163": "NON_PARTICIPATING",
"164": "NOT_ALIVE",
"165": "NOT_ALIVE",
"166": "NON_PARTICIPATING",
"167": "SUCCESS",
"168": "SUCCESS",
"169": "NON_PARTICIPATING",
"17": "NOT_ALIVE",
"170": "SUCCESS",
"171": "NON_PARTICIPATING",
"172": "NOT_ALIVE",
"173": "NON_PARTICIPATING",
"174": "NOT_ALIVE",
"175": "SUCCESS",
"176": "NOT_ALIVE",
"177": "NON_PARTICIPATING",
"178": "NON_PARTICIPATING",
"179": "NOT_ALIVE",
"18": "NOT_ALIVE",
"180": "NOT_ALIVE",
"181": "NOT_ALIVE",
"182": "NON_PARTICIPATING",
"183": "NOT_ALIVE",
"184": "SUCCESS",
"185": "SUCCESS",
"186": "NOT_ALIVE",
"187": "SUCCESS",
"188": "SUCCESS",
"189": "NOT_ALIVE",
"19": "NON_PARTICIPATING",
"190": "NON_PARTICIPATING",
"191": "NOT_ALIVE",
"192": "NON_PARTICIPATING",
"193": "NOT_ALIVE",
"194": "NOT_ALIVE",
"195": "NOT_ALIVE",
"196": "SUCCESS",
"197": "NON_PARTICIPATING",
"198": "SUCCESS",
"199": "NON_PARTICIPATING",
"2": "NOT_ALIVE",
"20": "SUCCESS",
"200": "NOT_ALIVE",
"201": "NOT_ALIVE",
"202": "NON_PARTICIPATING",
"203": "NON_PARTICIPATING",
"204": "NOT_ALIVE",
"205": "SUCCESS",
"206": "NON_PARTICIPATING",
"207": "NOT_ALIVE",
"208": "NOT_ALIVE",
"209": "NON_PARTICIPATING",
"21": "NON_PARTICIPATING",
"210": "NON_PARTICIPATING",
"211": "NON_PARTICIPATING",
"212": "SUCCESS",
"213": "NOT_ALIVE",
"214": "NOT_ALIVE",
"215": "NOT_ALIVE",
"216": "NOT_ALIVE",
"217": "NOT_ALIVE",
"218": "NOT_ALIVE",
"219": "NON_PARTICIPATING",
"22": "NOT_ALIVE",
"220": "NON_PARTICIPATING",
"221": "SUCCESS",
"222": "NON_PARTICIPATING",
"223": "NON_PARTICIPATING",
"224": "NOT_ALIVE",
"225": "NOT_ALIVE",
"226": "NON_PARTICIPATING",
"227": "SUCCESS",
"228": "NOT_ALIVE",
"229": "NOT_ALIVE",
"23": "NOT_ALIVE",
"230": "NON_PARTICIPATING",
"231": "SUCCESS",
"232": "NOT_ALIVE",
"233": "NOT_ALIVE",
"234": "NON_PARTICIPATING",
"235": "NON_PARTICIPATING",
"236": "NOT_ALIVE",
"237": "SUCCESS",
"238": "NON_PARTICIPATING",
"239": "NON_PARTICIPATING",
"24": "NON_PARTICIPATING",
"240": "NON_PARTICIPATING",
"241": "NON_PARTICIPATING",
"242": "NOT_ALIVE",
"243": "NOT_ALIVE",
"244": "NOT_ALIVE",
"245": "NON_PARTICIPATING",
"246": "SUCCESS",
"247": "NOT_ALIVE",
"248": "FAIL",
"249": "NOT_ALIVE",
"25": "NOT_ALIVE",
"250": "SUCCESS",
"251": "NOT_ALIVE",
"252": "NON_PARTICIPATING",
"253": "NOT_ALIVE",
"254": "NON_PARTICIPATING",
"255": "SUCCESS",
"26": "NON_PARTICIPATING",
"27": "NOT_ALIVE",
"28": "NON_PARTICIPATING",
"29": "NON_PARTICIPATING",
"3": "NOT_ALIVE",
"30": "NON_PARTICIPATING",
"31": "NOT_ALIVE",
"32": "NOT_ALIVE",
"33": "NOT_ALIVE",
"34": "NON_PARTICIPATING",
"35": "SUCCESS",
"36": "NON_PARTICIPATING",
"37": "NON_PARTICIPATING",
"38": "NON_PARTICIPATING",
"39": "NON_PARTICIPATING",
"4": "SUCCESS",
"40": "NON_PARTICIPATING",
"41": "NOT_ALIVE",
"42": "NON_PARTICIPATING",
"43": "NOT_ALIVE",
"44": "SUCCESS",
"45": "NON_PARTICIPATING",
"46": "NON_PARTICIPATING",
"47": "NOT_ALIVE",
"48": "NOT_ALIVE",
"49": "NOT_ALIVE",
"5": "NOT_ALIVE",
"50": "NON_PARTICIPATING",
"51": "NOT_ALIVE",
"52": "NOT_ALIVE",
"53": "NON_PARTICIPATING",
"54": "NOT_ALIVE",
"55": "NOT_ALIVE",
"56": "NON_PARTICIPATING",
"57": "FAIL",
"58": "NOT_ALIVE",
"59": "SUCCESS",
"6": "NON_PARTICIPATING",
"60": "FAIL",
"61": "SUCCESS",
"62": "NON_PARTICIPATING",
"63": "NON_PARTICIPATING",
"64": "NOT_ALIVE",
"65": "NON_PARTICIPATING",
"66": "SUCCESS",
"67": "NON_PARTICIPATING",
"68": "NON_PARTICIPATING",
"69": "NON_PARTICIPATING",
"7": "SUCCESS",
"70": "NOT_ALIVE",
"71": "NOT_ALIVE",
"72": "NON_PARTICIPATING",
"73": "NON_PARTICIPATING",
"74": "SUCCESS",
"75": "NON_PARTICIPATING",
"76": "NOT_ALIVE",
"77": "NOT_ALIVE",
"78": "NOT_ALIVE",
"79": "NON_PARTICIPATING",
"8": "SUCCESS",
"80": "NON_PARTICIPATING",
"81": "NOT_ALIVE",
"82": "SUCCESS",
"83": "NOT_ALIVE",
"84": "SUCCESS",
"85": "NOT_ALIVE",
"86": "SUCCESS",
"87": "SUCCESS",
"88": "NOT_ALIVE",
"89": "SUCCESS",
"9": "NON_PARTICIPATING",
"90": "NOT_ALIVE",
"91": "SUCCESS",
"92": "SUCCESS",
"93": "NOT_ALIVE",
"94": "SUCCESS",
"95": "NOT_ALIVE",
"96": "SUCCESS",
"97": "NOT_ALIVE",
"98": "NOT_ALIVE",
"99": "FAIL"
},
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"block_list": [
6512144,
6512180
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"inner_step": 40,
"intermediate_size": 5632,
"last_allreduce_block": 6510383,
"max_position_embeddings": 2048,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 22,
"num_key_value_heads": 4,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.39.3",
"use_cache": false,
"vocab_size": 32000
}