a6 / config.json
zeras141a's picture
Run 8. Outer Step 128. Inner Step 91.
9708176 verified
{
"_name_or_path": "distributed/llama-1b",
"all_reduce_scores": {
"0": "NOT_ALIVE",
"1": "NOT_ALIVE",
"10": "NOT_ALIVE",
"100": "NOT_ALIVE",
"101": "NOT_ALIVE",
"102": "NOT_ALIVE",
"103": "NON_PARTICIPATING",
"104": "NOT_ALIVE",
"105": "NOT_ALIVE",
"106": "NOT_ALIVE",
"107": "NON_PARTICIPATING",
"108": "NON_PARTICIPATING",
"109": "SUCCESS",
"11": "NOT_ALIVE",
"110": "NON_PARTICIPATING",
"111": "SUCCESS",
"112": "SUCCESS",
"113": "NON_PARTICIPATING",
"114": "SUCCESS",
"115": "NOT_ALIVE",
"116": "NON_PARTICIPATING",
"117": "NON_PARTICIPATING",
"118": "NON_PARTICIPATING",
"119": "NON_PARTICIPATING",
"12": "NOT_ALIVE",
"120": "NOT_ALIVE",
"121": "NON_PARTICIPATING",
"122": "NON_PARTICIPATING",
"123": "SUCCESS",
"124": "NON_PARTICIPATING",
"125": "NOT_ALIVE",
"126": "NOT_ALIVE",
"127": "SUCCESS",
"128": "NON_PARTICIPATING",
"129": "NON_PARTICIPATING",
"13": "SUCCESS",
"130": "NON_PARTICIPATING",
"131": "NON_PARTICIPATING",
"132": "NON_PARTICIPATING",
"133": "SUCCESS",
"134": "NON_PARTICIPATING",
"135": "NOT_ALIVE",
"136": "NON_PARTICIPATING",
"137": "NON_PARTICIPATING",
"138": "NON_PARTICIPATING",
"139": "NON_PARTICIPATING",
"14": "NOT_ALIVE",
"140": "NON_PARTICIPATING",
"141": "SUCCESS",
"142": "SUCCESS",
"143": "NON_PARTICIPATING",
"144": "NON_PARTICIPATING",
"145": "NOT_ALIVE",
"146": "SUCCESS",
"147": "NON_PARTICIPATING",
"148": "NON_PARTICIPATING",
"149": "NON_PARTICIPATING",
"15": "NOT_ALIVE",
"150": "NOT_ALIVE",
"151": "NON_PARTICIPATING",
"152": "SUCCESS",
"153": "NON_PARTICIPATING",
"154": "NOT_ALIVE",
"155": "NOT_ALIVE",
"156": "NOT_ALIVE",
"157": "NOT_ALIVE",
"158": "SUCCESS",
"159": "NOT_ALIVE",
"16": "NOT_ALIVE",
"160": "NON_PARTICIPATING",
"161": "NOT_ALIVE",
"162": "NON_PARTICIPATING",
"163": "NOT_ALIVE",
"164": "NON_PARTICIPATING",
"165": "SUCCESS",
"166": "NON_PARTICIPATING",
"167": "SUCCESS",
"168": "NON_PARTICIPATING",
"169": "NOT_ALIVE",
"17": "NOT_ALIVE",
"170": "NOT_ALIVE",
"171": "NON_PARTICIPATING",
"172": "NON_PARTICIPATING",
"173": "SUCCESS",
"174": "NOT_ALIVE",
"175": "SUCCESS",
"176": "NOT_ALIVE",
"177": "NON_PARTICIPATING",
"178": "SUCCESS",
"179": "NON_PARTICIPATING",
"18": "NOT_ALIVE",
"180": "NOT_ALIVE",
"181": "NON_PARTICIPATING",
"182": "SUCCESS",
"183": "SUCCESS",
"184": "NON_PARTICIPATING",
"185": "SUCCESS",
"186": "NOT_ALIVE",
"187": "NON_PARTICIPATING",
"188": "NON_PARTICIPATING",
"189": "SUCCESS",
"19": "NOT_ALIVE",
"190": "SUCCESS",
"191": "NON_PARTICIPATING",
"192": "NON_PARTICIPATING",
"193": "NON_PARTICIPATING",
"194": "NON_PARTICIPATING",
"195": "SUCCESS",
"196": "SUCCESS",
"197": "NOT_ALIVE",
"198": "SUCCESS",
"199": "SUCCESS",
"2": "NOT_ALIVE",
"20": "NOT_ALIVE",
"200": "NON_PARTICIPATING",
"201": "SUCCESS",
"202": "SUCCESS",
"203": "SUCCESS",
"204": "NOT_ALIVE",
"205": "NON_PARTICIPATING",
"206": "SUCCESS",
"207": "NON_PARTICIPATING",
"208": "NON_PARTICIPATING",
"209": "SUCCESS",
"21": "NOT_ALIVE",
"210": "SUCCESS",
"211": "NON_PARTICIPATING",
"212": "NON_PARTICIPATING",
"213": "NON_PARTICIPATING",
"214": "NON_PARTICIPATING",
"215": "NON_PARTICIPATING",
"216": "SUCCESS",
"217": "NOT_ALIVE",
"218": "SUCCESS",
"219": "NOT_ALIVE",
"22": "NOT_ALIVE",
"220": "NON_PARTICIPATING",
"221": "NON_PARTICIPATING",
"222": "NON_PARTICIPATING",
"223": "NON_PARTICIPATING",
"224": "NON_PARTICIPATING",
"225": "SUCCESS",
"226": "NON_PARTICIPATING",
"227": "NON_PARTICIPATING",
"228": "SUCCESS",
"229": "NOT_ALIVE",
"23": "NOT_ALIVE",
"230": "NON_PARTICIPATING",
"231": "NON_PARTICIPATING",
"232": "NOT_ALIVE",
"233": "NON_PARTICIPATING",
"234": "NON_PARTICIPATING",
"235": "NON_PARTICIPATING",
"236": "SUCCESS",
"237": "SUCCESS",
"238": "NOT_ALIVE",
"239": "NON_PARTICIPATING",
"24": "NOT_ALIVE",
"240": "NON_PARTICIPATING",
"241": "NON_PARTICIPATING",
"242": "NOT_ALIVE",
"243": "NON_PARTICIPATING",
"244": "NON_PARTICIPATING",
"245": "SUCCESS",
"246": "NON_PARTICIPATING",
"247": "SUCCESS",
"248": "NON_PARTICIPATING",
"249": "SUCCESS",
"25": "SUCCESS",
"250": "NON_PARTICIPATING",
"251": "NOT_ALIVE",
"252": "NOT_ALIVE",
"253": "SUCCESS",
"254": "NOT_ALIVE",
"255": "NOT_ALIVE",
"26": "NOT_ALIVE",
"27": "NON_PARTICIPATING",
"28": "NON_PARTICIPATING",
"29": "NON_PARTICIPATING",
"3": "NOT_ALIVE",
"30": "NON_PARTICIPATING",
"31": "NON_PARTICIPATING",
"32": "NON_PARTICIPATING",
"33": "NOT_ALIVE",
"34": "NON_PARTICIPATING",
"35": "NON_PARTICIPATING",
"36": "NON_PARTICIPATING",
"37": "NON_PARTICIPATING",
"38": "NON_PARTICIPATING",
"39": "NOT_ALIVE",
"4": "NOT_ALIVE",
"40": "SUCCESS",
"41": "NON_PARTICIPATING",
"42": "NON_PARTICIPATING",
"43": "SUCCESS",
"44": "NOT_ALIVE",
"45": "NON_PARTICIPATING",
"46": "NON_PARTICIPATING",
"47": "SUCCESS",
"48": "NON_PARTICIPATING",
"49": "SUCCESS",
"5": "NOT_ALIVE",
"50": "NOT_ALIVE",
"51": "SUCCESS",
"52": "SUCCESS",
"53": "NON_PARTICIPATING",
"54": "NON_PARTICIPATING",
"55": "NON_PARTICIPATING",
"56": "NON_PARTICIPATING",
"57": "NOT_ALIVE",
"58": "NOT_ALIVE",
"59": "NOT_ALIVE",
"6": "SUCCESS",
"60": "SUCCESS",
"61": "SUCCESS",
"62": "NOT_ALIVE",
"63": "SUCCESS",
"64": "NON_PARTICIPATING",
"65": "SUCCESS",
"66": "NOT_ALIVE",
"67": "NON_PARTICIPATING",
"68": "SUCCESS",
"69": "NON_PARTICIPATING",
"7": "NOT_ALIVE",
"70": "NON_PARTICIPATING",
"71": "NON_PARTICIPATING",
"72": "NON_PARTICIPATING",
"73": "NON_PARTICIPATING",
"74": "SUCCESS",
"75": "NON_PARTICIPATING",
"76": "NON_PARTICIPATING",
"77": "NON_PARTICIPATING",
"78": "NON_PARTICIPATING",
"79": "NOT_ALIVE",
"8": "NOT_ALIVE",
"80": "NOT_ALIVE",
"81": "NON_PARTICIPATING",
"82": "NON_PARTICIPATING",
"83": "NOT_ALIVE",
"84": "NON_PARTICIPATING",
"85": "NON_PARTICIPATING",
"86": "NON_PARTICIPATING",
"87": "NON_PARTICIPATING",
"88": "NON_PARTICIPATING",
"89": "NON_PARTICIPATING",
"9": "NOT_ALIVE",
"90": "NON_PARTICIPATING",
"91": "NON_PARTICIPATING",
"92": "NON_PARTICIPATING",
"93": "NON_PARTICIPATING",
"94": "SUCCESS",
"95": "NON_PARTICIPATING",
"96": "SUCCESS",
"97": "SUCCESS",
"98": "FAIL",
"99": "SUCCESS"
},
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"block_list": [
6445126,
6445140
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"inner_step": 90,
"intermediate_size": 5632,
"last_allreduce_block": 6444495,
"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
}