a4 / config.json
zeras141a's picture
Run 8. Outer Step 72. Inner Step 53.
13bfd9a verified
raw
history blame
6.95 kB
{
"_name_or_path": "distributed/llama-1b",
"all_reduce_scores": {
"0": "NOT_ALIVE",
"1": "NOT_ALIVE",
"10": "NOT_ALIVE",
"100": "NON_PARTICIPATING",
"101": "FAIL",
"102": "FAIL",
"103": "NOT_ALIVE",
"104": "NON_PARTICIPATING",
"105": "NON_PARTICIPATING",
"106": "FAIL",
"107": "FAIL",
"108": "NON_PARTICIPATING",
"109": "NON_PARTICIPATING",
"11": "NOT_ALIVE",
"110": "NOT_ALIVE",
"111": "FAIL",
"112": "FAIL",
"113": "NOT_ALIVE",
"114": "NON_PARTICIPATING",
"115": "FAIL",
"116": "NON_PARTICIPATING",
"117": "FAIL",
"118": "NON_PARTICIPATING",
"119": "FAIL",
"12": "NOT_ALIVE",
"120": "NON_PARTICIPATING",
"121": "FAIL",
"122": "FAIL",
"123": "FAIL",
"124": "FAIL",
"125": "NOT_ALIVE",
"126": "FAIL",
"127": "FAIL",
"128": "FAIL",
"129": "FAIL",
"13": "FAIL",
"130": "FAIL",
"131": "NON_PARTICIPATING",
"132": "FAIL",
"133": "FAIL",
"134": "FAIL",
"135": "NON_PARTICIPATING",
"136": "FAIL",
"137": "NON_PARTICIPATING",
"138": "FAIL",
"139": "FAIL",
"14": "NOT_ALIVE",
"140": "FAIL",
"141": "NON_PARTICIPATING",
"142": "FAIL",
"143": "FAIL",
"144": "NON_PARTICIPATING",
"145": "FAIL",
"146": "NON_PARTICIPATING",
"147": "NON_PARTICIPATING",
"148": "NON_PARTICIPATING",
"149": "FAIL",
"15": "NOT_ALIVE",
"150": "FAIL",
"151": "NON_PARTICIPATING",
"152": "SUCCESS",
"153": "NON_PARTICIPATING",
"154": "FAIL",
"155": "FAIL",
"156": "NOT_ALIVE",
"157": "NON_PARTICIPATING",
"158": "NON_PARTICIPATING",
"159": "NON_PARTICIPATING",
"16": "FAIL",
"160": "NON_PARTICIPATING",
"161": "FAIL",
"162": "FAIL",
"163": "NON_PARTICIPATING",
"164": "FAIL",
"165": "FAIL",
"166": "NON_PARTICIPATING",
"167": "FAIL",
"168": "FAIL",
"169": "FAIL",
"17": "NOT_ALIVE",
"170": "NON_PARTICIPATING",
"171": "NON_PARTICIPATING",
"172": "NON_PARTICIPATING",
"173": "NON_PARTICIPATING",
"174": "FAIL",
"175": "FAIL",
"176": "NOT_ALIVE",
"177": "FAIL",
"178": "NOT_ALIVE",
"179": "NON_PARTICIPATING",
"18": "NOT_ALIVE",
"180": "FAIL",
"181": "NON_PARTICIPATING",
"182": "FAIL",
"183": "FAIL",
"184": "FAIL",
"185": "NOT_ALIVE",
"186": "FAIL",
"187": "NON_PARTICIPATING",
"188": "FAIL",
"189": "FAIL",
"19": "FAIL",
"190": "NON_PARTICIPATING",
"191": "FAIL",
"192": "NON_PARTICIPATING",
"193": "FAIL",
"194": "NON_PARTICIPATING",
"195": "FAIL",
"196": "FAIL",
"197": "NOT_ALIVE",
"198": "NOT_ALIVE",
"199": "FAIL",
"2": "NOT_ALIVE",
"20": "FAIL",
"200": "NON_PARTICIPATING",
"201": "NOT_ALIVE",
"202": "NON_PARTICIPATING",
"203": "NON_PARTICIPATING",
"204": "FAIL",
"205": "FAIL",
"206": "NON_PARTICIPATING",
"207": "NON_PARTICIPATING",
"208": "NON_PARTICIPATING",
"209": "FAIL",
"21": "NOT_ALIVE",
"210": "NOT_ALIVE",
"211": "NON_PARTICIPATING",
"212": "FAIL",
"213": "FAIL",
"214": "FAIL",
"215": "FAIL",
"216": "FAIL",
"217": "FAIL",
"218": "FAIL",
"219": "NON_PARTICIPATING",
"22": "FAIL",
"220": "NON_PARTICIPATING",
"221": "NON_PARTICIPATING",
"222": "NON_PARTICIPATING",
"223": "NON_PARTICIPATING",
"224": "NON_PARTICIPATING",
"225": "FAIL",
"226": "NON_PARTICIPATING",
"227": "NOT_ALIVE",
"228": "NOT_ALIVE",
"229": "NON_PARTICIPATING",
"23": "NOT_ALIVE",
"230": "NON_PARTICIPATING",
"231": "NON_PARTICIPATING",
"232": "NOT_ALIVE",
"233": "FAIL",
"234": "FAIL",
"235": "NON_PARTICIPATING",
"236": "NOT_ALIVE",
"237": "NON_PARTICIPATING",
"238": "NON_PARTICIPATING",
"239": "NON_PARTICIPATING",
"24": "FAIL",
"240": "FAIL",
"241": "NON_PARTICIPATING",
"242": "FAIL",
"243": "NON_PARTICIPATING",
"244": "FAIL",
"245": "FAIL",
"246": "NON_PARTICIPATING",
"247": "FAIL",
"248": "FAIL",
"249": "NOT_ALIVE",
"25": "SUCCESS",
"250": "FAIL",
"251": "FAIL",
"252": "NON_PARTICIPATING",
"253": "FAIL",
"254": "NOT_ALIVE",
"255": "NOT_ALIVE",
"26": "FAIL",
"27": "FAIL",
"28": "FAIL",
"29": "FAIL",
"3": "NOT_ALIVE",
"30": "FAIL",
"31": "NON_PARTICIPATING",
"32": "FAIL",
"33": "NOT_ALIVE",
"34": "FAIL",
"35": "FAIL",
"36": "FAIL",
"37": "FAIL",
"38": "NON_PARTICIPATING",
"39": "NON_PARTICIPATING",
"4": "NOT_ALIVE",
"40": "FAIL",
"41": "FAIL",
"42": "NON_PARTICIPATING",
"43": "FAIL",
"44": "FAIL",
"45": "NON_PARTICIPATING",
"46": "FAIL",
"47": "NON_PARTICIPATING",
"48": "FAIL",
"49": "FAIL",
"5": "NOT_ALIVE",
"50": "NOT_ALIVE",
"51": "NON_PARTICIPATING",
"52": "FAIL",
"53": "FAIL",
"54": "NON_PARTICIPATING",
"55": "FAIL",
"56": "NON_PARTICIPATING",
"57": "FAIL",
"58": "FAIL",
"59": "NON_PARTICIPATING",
"6": "NOT_ALIVE",
"60": "FAIL",
"61": "NON_PARTICIPATING",
"62": "FAIL",
"63": "FAIL",
"64": "NOT_ALIVE",
"65": "NON_PARTICIPATING",
"66": "NON_PARTICIPATING",
"67": "NOT_ALIVE",
"68": "FAIL",
"69": "NON_PARTICIPATING",
"7": "FAIL",
"70": "NOT_ALIVE",
"71": "NON_PARTICIPATING",
"72": "FAIL",
"73": "NON_PARTICIPATING",
"74": "FAIL",
"75": "NON_PARTICIPATING",
"76": "FAIL",
"77": "FAIL",
"78": "NON_PARTICIPATING",
"79": "FAIL",
"8": "NOT_ALIVE",
"80": "NON_PARTICIPATING",
"81": "NON_PARTICIPATING",
"82": "FAIL",
"83": "FAIL",
"84": "NON_PARTICIPATING",
"85": "NON_PARTICIPATING",
"86": "NOT_ALIVE",
"87": "FAIL",
"88": "NON_PARTICIPATING",
"89": "NOT_ALIVE",
"9": "NOT_ALIVE",
"90": "FAIL",
"91": "NOT_ALIVE",
"92": "FAIL",
"93": "FAIL",
"94": "NON_PARTICIPATING",
"95": "FAIL",
"96": "FAIL",
"97": "FAIL",
"98": "NON_PARTICIPATING",
"99": "NON_PARTICIPATING"
},
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"block_list": [
6394875,
6394892,
6394909
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"inner_step": 52,
"intermediate_size": 5632,
"last_allreduce_block": 6394503,
"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
}